AMER. ZOOL., 29:85-102 (1989)

Caudal Fin Locomotion in Ray-finned Fishes:Historical and Functional Analyses1

GEORGE V. LAUDER

School of Biological Sciences, University of California,Irvine, California 92717

SYNOPSIS. Over the last 20 years, considerable progress has been made in quantifyingthe movement of the body during locomotion by aquatic vertebrates, and in defining therole of axial musculature in producing these kinematic patterns. Relatively little is known,however, about how specific internal structural features of the axial system in fishes affectbody kinematics, and how such structural and functional features have changed duringevolution. The major theme of this paper is that historical, phylogenetic patterns in theaxial musculoskeletal system need to be integrated with experimental and functional datain order to understand the design of the locomotor apparatus in vertebrates. To illustratethis proposition, the evolution of the tail in ray-finned fishes is presented as a case studyin phylogenetic and functional analysis of the vertebrate axial musculoskeletal system.Traditionally, the evolution of the tail in ray-finned fishes has been viewed as a transfor-mation from a primitively heterocercal (functionally asymmetrical) tail to a homocercaltail in which the axis of rotation during locomotion was vertical, generating a symmetricalthrust. Both phylogenetic and functional approaches are used to examine this hypothesis.Major osteological and myological features of the tail in ray-finned fishes are mapped ontoa phylogeny of ray-finned fishes to discern historical sequences of morphological changein the axial musculoskeletal system. A key event in locomotor evolution was the origin ofthe hypochordal longitudinalis muscle, the only intrinsic caudal muscle with a line ofaction at an appreciable angle to the body axis. This muscle originated prior to the originof a caudal skeleton bearing both hypaxial and epaxial fin ray supports. The hypochordalmuscle is proposed to be a key component of the axial musculoskeletal system that allowsmost fishes to modulate caudal function and decouples external morphological symmetryfrom functional symmetry. Experimental data (strain gauge recordings from tail bones,and electromyographic recordings from intrinsic and extrinsic caudal muscles) corroboratethis interpretation and suggest that functional symmetry in the tail of ray-finned fishes isnot predictable from skeletal morphology alone, but depends on the activity of the hypo-chordal longitudinalis muscle and on locomotor mode. The homocercal teleost tail maythus function asymmetrically.

INTRODUCTION aquatic propulsion. The analysis of fishThe study of aquatic locomotion has locomotion has served as a primary exam-

received more attention over the last 20 Pl e f o r t h e analysis of how environmentalyears than any other aspect of vertebrate demands constrain organismal designfunctional morphology. Numerous recent (Weihs, 1989). The dense and viscousbooks and symposia have dealt with the aquatic medium places severe constraintsmechanisms by which vertebrates move on the functional design of propulsive sys-through the water (e.g., Gray, 1968; Aleyev, terns, and both theoretical models of these1969; Webb, 1975; Wu et al., 1975; Blake, constraints and the experimental demon-1983). In part because of their extraordi- stration of design limitations have used ray-nary taxonomic and locomotor diversity, finned fishes as a model system,ray-finned fishes (Actinopterygii) have pro- There are three areas in which partic-vided key examples for the analysis of loco- ularly significant progress has been mademotor movements and data used to test in the analysis offish locomotion. First, thetheoretical hydrodynamic models of kinematics of aquatic propulsion has been

extensively studied and these data havebeen used to test theoretical models of bodymovement (Gray, 1933a, b; Bainbridge,

• From the Symposium on Axial Movement Systems: gQi u h h n , l 9 7 l ; W e i h S ) 1 9 7 3 ; W e bb ,Biomechanics and Neural Control p r e s e n t e d a t t h e ,nr,o , ? • > , r>o-<\ \r •Annual Meeting of the American Society of Zoolo- 1978a; Vldeler and HeSS, 1984). Major cat-gists, 27-30 December 1986, at Nashville, Tennessee. egories of locomotor mode have been

85

86 GEORGE V. LAUDER

defined on the basis of kinematic patternssuch as the well-known anguilliform, tun-niform, and ostraciiform locomotor classes(Breder, 1926; Webb, 1975; Lindsey,1978). Secondly, the role of the myotomalmusculature and skin fiber systems in pro-ducing body deformations has been thesubject of continuing study (Bone, 1966,1978, 1988; Johnston, 1980; Wainwright,1983) and the mechanisms by which mus-cle contractions act to bend the body havebeen analyzed. Thirdly, an important setof functional and design constraints havebeen identified by Webb (1978a, b). He hasshown that the functional requirements forhigh performance in steady, continuouslocomotion are in conflict with demandsfor high accelerations (unsteady locomo-tion).

Despite this progress in the area of bio-mechanics and functional morphology ofaquatic propulsion in fishes, there are majorconceptual avenues that have yet to beexplored. For example, almost no progresshas been made in synthesizing the resultsfrom biomechanical and functional studieswith historical, phylogenetic analyses. Inaddition, almost nothing is known aboutthe functional significance of morpholog-ical features in the axial musculoskeletalsystem that have been used by systematiststo define lineages of ray-finned fishes.

The major theme of this paper is thathistorical patterns in the axial musculo-skeletal system of fishes need to be inte-grated with experimental and functionaldata: there has been no synthesis of exper-imental and phylogenetic research so nec-essary to a general understanding of theevolution of structure and function in theaxial musculoskeletal system. Without asynthesis of historical and functional data,it will not be possible to understand thedesign of vertebrate locomotor systems.The aim of this essay is to illustrate thispoint and provide a case study of both func-tional and phylogenetic analysis of thelocomotor apparatus. It is beyond the scopeof the paper to consider all aspects of loco-motor structure and function. Rather, Iwill concentrate on presenting a case studyof a classical system: evolution of the tailin rav-finned fishes.

PHYLOGENY AND BIOMECHANICS OF THECAUDAL FIN: A CASE STUDY

Nearly every investigator who has dis-cussed the evolution of vertebrates hascommented on the caudal fin of ray-finnedfishes (Actinopterygii). Caudal fin evolu-tion has become a textbook case of struc-tural and functional modification in ver-tebrates, and is used to illustrate howchanges in external morphology haveoccurred and had important functionalconsequences (e.g., Goodrich, 1930;Jollie,1972; Romer and Parsons, 1986). Furtheremphasizing the importance of this system,many investigators have attributed thediversification of teleost fishes into over23,000 species, by far the largest clade ofvertebrates, to their locomotor abilities.Gosline (1971, p. 34) for example, notesthat "The perfection of caudal locomotionhas probably been the single greatestachievement of the teleostean fishes," andLund (1967, p. 216) comments that "theversatility of the teleostean caudal skeletonwas a very important factor in the rapidradiation of teleosts during the Mesozoic."

Exactly what are the changes that areproposed to have occurred in the acti-nopterygian caudal skeleton? Primitive ray-finned fishes possess a heterocercal tail inwhich the notochord and vertebral ele-ments extend into the upper (dorsal) taillobe (Fig. IB). The fin rays and the sup-porting skeletal elements thus attach to theventral surface of vertebral axis. This mor-phological arrangement is proposed to havegenerated an epibatic (lift) force as the tailwas swept from side to side (Affleck, 1950;Alexander, 1967; Aleyev, 1969). The het-erocercal tail morphology is viewed as bothstructurally and functionally asymmetricalin that the tail is formed primarily out ofhypaxial structures, and the direction ofthrust produced by the tail does not passthrough the vertebral axis. The inclinedangle of fin rays with respect to the verticaland the upturned vertebral axis defines anoblique line of fin ray oscillation (Fig. IB,C).

It is important at this point to clearlydistinguish between epaxial and hypaxialstructures in the caudal fin, and the terms

LOCOMOTION IN RAY-FINNED FISHES 87

FIG. 1. Figure of homocercal and heterocercal tails and lines of bending based on Affleck (1950). A-D. Fourdiagrams of different external tail morphologies showing some of the variation in tail shape within ray-finnedfishes. E-H. Diagrammatic sequence of tail ontogeny in a teleost fish. The dashed line indicates the line ofbending (BA) of the tail as proposed by Affleck (1950). The homocercal tail (D and H) is proposed to havea vertical line of bending. Shapes A, B, and C represent heterocercal tails with an oblique line of bending.The arrows indicate the directions and relative magnitudes of thrust produced by the dorsal and ventral taillobes as proposed by Affleck (1950). Note the distinction between the developing epaxial (EP) and hypaxial(HY) components of caudal structure on either side of the notochordal axis (NA) shown in F, and the dorsal(DOR) and ventral (VEN) lobes of the tail shown in H.

"dorsal" and "ventral" that are oftenapplied to caudal structures. Epaxial andhypaxial refer to structures that originatedevelopmentally from above and below thenotochordal axis respectively (Fig. 1;Whitehouse, 1910). These terms are notsynonymous with dorsal and ventral, whichrefer to the positions relative to a horizon-tal axis through the middle of the body.As fin rays in most fishes develop fromhypaxial tissues (Fig. 1), they are properlyreferred to as hypaxial structures. Only afew components of the caudal skeleton areepaxial in origin. Teleost fish tails, how-ever, have roughly equal dorsal and ventrallobes; some hypaxial fin rays thus articulatewith epaxial endoskeletal elements.

The homocercal tail of teleost fishes isexternally symmetrical (Fig. ID). Bothepaxial and hypaxial derivatives contributeto the internal caudal skeleton, and the axisof rotation is proposed to be vertical: finrays oscillate about a vertical axis wherethey join the skeletal supports within thetail. The functional result of the externalsymmetry and vertical axis of rotation ispurportedly that the direction of thrustproduced by the tail passes through thevertebral axis, and does not generate lift

forces (Aleyev, 1969). Olson (1971, p. 534)exemplifies this view: "Evolution in theactinopterygians proceeded through suc-cessive stages to a reduced heterocercal tail,found in subholosteans and some holos-teans, and to a strictly homocercal condi-tion in which the tail fin is symmetricallydisposed as in many teleosts. Functionally,the fin so produced is isobatic, with thethrust being entirely forward." It is impor-tant to note that this external symmetry isnot mirrored by the internal skeletal struc-ture: epaxial and hypaxial components ofthe tail are not mirror images, but theasymmetrical caudal skeleton does supportexternally symmetrical fin rays.

Considerable significance has beenattached in the literature to the evolution-ary transformation within ray-finned fishesfrom the primitive heterocercal conditionto the teleost homocercal shape (Patterson,1968; Nybelin, 1973), and one of the dom-inant arguments is that the homocercal tailis more efficient than the heterocercal tail.As Affleck (1950, p. 365) comments,"Because the fin of a homocercal tail swingsabout a vertical axis it is more efficient aspart of the propulsive unit than the fins ofa heterocercal tail." Patterson (1968, p.

88 GEORGE V. LAUDER

233) proposed that teleost fishes had anadvantage in locomotion because of "theincreased efficiency in horizontal swim-ming of a fish in which both lobes of thetail are equal in area and in flexibility, inwhich the axis is not upturned, and in whichthe tail swings about a vertical rather thanan oblique axis . . . ." Gosline (1971, pp.35—36) concurs, stating that a teleost fish"can swing both caudal lobes back and forthsynchronously, both lobes generating adirectly forward force."

It is critical to note several features ofthe proposed differences between hetero-cercal and homocercal tails. First, actualfunctional differences between these twotail types in ray-finned fishes have not beendemonstrated in vivo. Secondly, the precisesequence of structural modification of theheterocercal tail into the homocercal tailhas not been documented. Thirdly, vir-tually all analyses have focused on the skel-etal system, and have ignored the intrinsiccaudal muscles that might function to mod-ify the position of the fin rays and caudalskeleton. And, fourthly, all workers havediscussed caudal function in ray-finnedfishes during continuous locomotion, andhave not addressed the possibility that cau-dal morphology may reflect functionaldemands imposed by both continuous loco-motion and fast-start accelerations.

The purpose of the case study of caudalevolution in ray-finned fishes presentedbelow is to address these unresolved issuesin more detail. Specifically the followingquestions will be considered. (1) What isthe precise historical sequence of modifi-cation of the caudal skeleton and muscu-lature in ray-finned fishes? (2) Does thehomocercal tail of teleosts actually func-tion symmetrically as has been assumed? Inother words, does external symmetry implyfunctional symmetry? (3) Does the functionof the homocercal teleost tail changebetween continuous locomotion and fast-start accelerations?

PHYLOGENETIC PATTERNS TOCAUDAL STRUCTURE

Historical analysis

By what specific historical sequence wasthe functional design of the homocercal

teleost tail produced? Did modifications tothe epaxial portion of the tail occur priorto those in the hypaxial portion, and howdid the ability of teleost fishes to alter thearea of the caudal fin evolve? Historicalsequences such as these can be resolved bymapping structural features of the caudalmusculoskeletal system onto a previouslyestablished phylogeny. Figure 2 summa-rizes the results of such a procedure, andthe phylogeny used is based on work sum-marized by Lauder and Liem (1983). Thefollowing discussion will analyze thesequence by which musculoskeletal novel-ties in the caudal skeleton and musculaturewere acquired at successive phylogeneticlevels, starting with the basal (primitive)condition for ray-finned fishes. In this way,an explicit historical sequence can bereconstructed by which the complex designof the homocercal tail was built up.

Primitive ray-finned fishes such as Chei-rolepis (Pearson, 1982) and Moythomasia(Gardiner, 1984) had a strongly hetero-cercal tail in which the hypaxial lobeappears to have had no intrinsic muscula-ture. There is no contribution of the epax-ial skeleton or myotomes to caudal finstructure. Based on the attachment of themost superficial lateral myotomal fibers(often designated as the lateralis superfi-cialis muscle, a portion of which is shownin Fig. 8) to the head of the fin rays inalmost all extant ray-finned fishes, it is par-simonious to assume that lateralis superfi-cialis in early ray-finned fishes also attachedto the heads of the fin rays and was capableof affecting at least minor changes in finray position. The tail of Polypterus, the mostprimitive living actinopterygian (Figs. 2, 3),is specialized with respect to other lowerray-finned fishes in having a diphycercalmorphology with both epaxial and hypax-ial lobes. Dissections of Polypterus revealthat the lateral myotomes thin posteriorlyto attach to the heads of the caudal fin rays(Fig. 3). There is no differentiation ofsuperficial and deep caudal musculature inPolypterus and no separation of distinctintrinsic muscles.

In the Chondrostei (Figs. 2, 4), numer-ous hypaxial fin rays attach to small carti-lages ventral to the notochord, and no

LOCOMOTION IN RAY-FINNED FISHES 89

Node 6Node 5

Node 4

Node 7

Node 3

Node 2

Node 1

Fie. 2. Phylogenetic relationships of the major clades of ray-finned fishes (Actinopterygii). Taxa markedwith a " + " contain no living species. This phylogeny is based on work summarized in Lauder and Liem(1983), notably Patterson (1973, 1977, 1982), Patterson and Rosen (1977), Wiley (1976), and Schaeffer andMcDonald (1978), as well as Gardiner (1984). Taxa derived from Node 1 = Actinopterygii; taxa derived fromNode 3 = Halecostomi; taxa derived from Node 4 = Teleostei. Using this phylogeny as a basis, morphologicalnovelties in the axial musculoskeletal system were identified at successive phylogenetic levels. Those featuresprimitive for ray-finned fishes, and thus present at Node 1 are: the absence of intrinsic caudal musculature,lateralis superficialis muscle attached to the heads of the caudal fin rays, heterocercal tail, posterior haemalspines directed posteroventrally. Locomotor characters present at the other nodes on the cladogram are:Node 2: hypurals on posterior caudal vertebrae oriented anteroposteriorly, epurals present (Schultze andArratia, 1986) but not supporting dorsal fin rays, weakly developed hypochordal longitudinalis muscle, distinctsuperficial and deep intrinsic caudal muscle layers, large flexor ventralis muscle; Node 3: epaxial interradialisand supracarinalis posterior muscles present, well-developed hypochordal longitudinalis; Node 4: ural neuralarches modified into uroneurals (but not expanded into supporting elements for dorsal fin rays), expansionof posteriorly directed haemal spines; Node 5: caudal skeleton has only two ural centra (vs. three or moreprimitively), expanded uroneurals, posterodorsally located epurals (epaxial elements) serving to support dorsalfin rays; Node 6: seven hypurals (vs. eight or more primitively), hypochordal longitudinalis muscle originatesfrom the ventral hypurals (and not from the vertebral column), hypaxial interradialis and infracarinalisposterior muscles present; Node 7: only two uroneurals extend anteriorly past second ural centrum.

intrinsic caudal muscles are present. Ten-dons from the posterior myotomes extendposterodorsally along the notochord andattach to connective tissue overlying thenotochord and tail cartilages.

In the Ginglymodi (gars and their rela-tives; Wiley, 1976) there are no epaxialsupports for caudal rays, but well devel-oped hypural bones (modified haemalspines) are present and support the caudalfin rays (Fig. 5). At this phylogenetic level(Fig. 2: Node 2) distinct superficial and deepintrinsic caudal musculature is present.Gars also have a weakly developed hypo-chordal longitudinalis muscle (Fig. 5: HL).This muscle originates on the caudal ver-tebrae and extends posterodorsally to inserton the first dorsal fin ray. This muscle is

the first intrinsic caudal muscle to differ-entiate phylogenetically in ray-finned fishesand it occurs in clades retaining the prim-itive heterocercal tail condition (Fig. 2:between Nodes 2 and 3). The hypochordallongitudinalis in gars is a differentiateddivision of a large intrinsic ventral flexormuscle which extends broadly from thevertebral column posteroventrally to inserton and between the heads of the fin rays(Fig. 5: FV). The ventral flexor musclespans the entire lateral surface of thehypural bones between the vertebral col-umn and fin ray heads, and is subdividedinto many bundles running nearly parallelto the hypurals; the hypochordal longitu-dinalis is merely a particularly well differ-entiated dorsal component of the flexor

10 GEORGE V. LAUDER

LSFIG. 3. Caudal skeleton and musculature of Polypterus senegalus. In this and Figures 4, 5, and 6, black linesindicate major muscles and their lines of action. Note the specialized symmetrical tail with caudal fin raysbeing supported by both epaxial and hypaxial skeletal elements. Polypterus lacks intrinsic caudal musculatureand shares with other primitive ray-finned fishes the condition of having the lateralis superficialis (LS) myoto-mal fibers attaching to the heads of the fin rays (black arrows).

ventralis. There is considerable musclefiber exchange between the muscle bun-dles composing the flexor ventralis, whilethe hypochordal longitudinalis is separatedfrom the flexor ventralis by a distinct fas-cial plane. In gars, the hypochordal lon-gitudinalis forms an angle of about 30° withthe horizontal precaudal body axis, similarto the angle formed by the ventral fiberbundles of the flexor ventralis (Fig. 5: HL,FV). Superficial to the flexor ventralis is abroad lateralis superficialis muscle attach-ing to the heads of the caudal fin rays.

The Halecostomi (Fig. 2: Node 3) ofwhich the Halecomorphi (Amia and fossilrelatives) is the primitive sister clade, pos-

sess several morphological novelties in thecaudal region. The hypochordal longitu-dinalis muscle is well developed (Fig. 6) anddifferentiated from the surrounding super-ficial and deep intrinsic caudal muscula-ture. The hypochordal longitudinalis liesat an increased angle to the vertebral col-umn relative to the ginglymod condition,inserting on the dorsal fin rays at about a75° angle. The hypochordal longitudinalisoriginates from the posterior caudal ver-tebrae and narrows posterodorsally to ten-dons which insert on the first three dorsalfin rays (Fig. 6: HL). As in gars, there is alarge deep flexor ventralis that covers thehypurals laterally and is subdivided into

FIG. 4. Caudal skeleton and musculature of Acipenser stellatus. Note that there are no intrinsic caudal muscles.The lateral body myotomes condense posteriorly to a series of long tendons that run along the notochord(black arrow). The fin rays in this heterocercal tail are hypaxial structures.

LOCOMOTION IN RAY-FINNED FISHES 91

LS

FIG. 5. Caudal skeleton and musculature of Lepisosteus oculatus. Note the lateralis superficialis fibers (LS)attaching to the fin rays (arrows) and the extensive flexor ventralis (FV). Selected fiber bundles of the flexorventralis are diagrammatically indicated by black lines. The hypochordal longitudinalis (HL) is a distinctdivision of the flexor extending from the vertebral column to the dorsal fin rays.

distinct bundles. Lateralis superficialismuscle fibers attach to the fin ray headsand cover the flexor ventralis in lateral view(Fig. 6: LS). The morphology of the hypo-chordal longitudinalis muscle in Lepisosteusand Amia indicates that it differentiatedfrom the flexor ventralis, and not directlyfrom the myotomes as has been suggested(Winterbottom, 1974, p. 290).

Two related novelties in the muscula-ture of the dorsal portion of the caudal finalso occur at this phylogenetic level (Fig.2: Node 3). Interradialis muscles are pres-ent between the dorsal fin rays and a supra-carinalis posterior muscle (Winterbottom,1974) extends from the posterior base ofthe dorsal fin to attach to the upper caudalfin rays (Fig. 6: IR, SCP). These two setsof muscles are antagonistic to each otherand allow expansion, stabilization, andcontraction of the dorsal caudal fin rays;caudal fin area can thus be modified. Thereare no ventral interradialis or infracari-nalis posterior muscles present at this phy-logenetic level.

At the next phylogenetic level (Fig. 2:Node 4), a major innovation occurs in thecaudal skeleton. As first adduced by Pat-

terson (1968), the Teleostei are defined bythe presence of elongate ural (tail) neuralarches called uroneurals. These elementsprovide the first major stiffening compo-nent of the epaxial portion of the caudalskeleton, but at this hierarchical level donot support fin rays posteriorly.

Significant expansion of the epaxial sup-port in the tail occurs at the next phylo-genetic level within the Teleostei (Fig. 2:Node 5; Table 1). The uroneurals of ich-thyodectiform fishes are expanded to formsupports for the dorsal fin rays, and epuralbones provide added stiffening of the uppercaudal lobe (Patterson, 1968, 1973). Therehas also been a reduction in the numberof vertebral centra in the tail region at thislevel.

Living teleost clades (Fig. 2: Node 6)share further modifications of the intrinsiccaudal musculature not found in Amia (Figs.7, 8). Both epaxial and hypaxial carinalmuscles are present, and both epaxial andhypaxial interradialis muscles occur. Thehypochordal longitudinalis has a new ori-gin from ventral hypural bones (not fromthe vertebral column), and this conditionis maintained within the vast majority of

m GEORGE V. LAUDER

FIG. 6. Caudal skeleton and musculature of Amia calva. Note the interradialis muscles (IR) on the dorsal finrays only, and the supracarinalis posterior (SCP) attaching to the upper fin rays. An extensive deep flexorventralis (FV) is present as is the lateralis superficialis (LS).

teleost fishes (Fig. 8). The hypochordallongitudinalis is a very conservative musclephylogenetically, showing relatively littlevariation within teleosts and maintainingits origin from the hypurals and its tendi-

nous insertion onto the most dorsal threeor four fin rays in taxa as divergent asosteoglossomorphs, ostariophysans, andpercomorphs (Nursall, 1963; Cowan, 1969;Liem, 1970; Winterbottom, 1974; Lauder,

T A B L E 1. Evolution of function in the caudal musculoskeletal system of ray-finned fishes. Nodes in the first column referto Figure 2. See text for discussion.

Phylogenetic level Major proposed functional characters

Number ofintrinsiccaudal

muscles*

Actinopterygii (Node 1)Ginglymodi + Halecos-

tomi (Node 2)

Halecostomi (Node 3)

Teleostei (Node 4)Ichthyodectiformes and

living teleost clades(Node 5)

Living teleost clades(Node 6)

Caudal fin ray movement coupled to myotomal contractions 0Lateral bending of caudal fin rays independent of myotomal con- 2

tractionLess muscular control of dorsal than ventral caudal fin rays in the

medio-lateral planeAbility to modulate caudal area in the sagittal plane using dorsal 4

caudal raysIncreased muscular control of dorsal fin rays via the hypochordal

longitudinalisStiffening of epaxial portion of the caudal skeleton 7Further stiffening and support for the epaxial portion of the caudal 7 +

skeleton

Control over caudal fin area complete, with acquisition of ability to 7 +move both dorsal and ventral fin rays in the sagittal plane

Ability to modulate dorsal fin ray stiffness increased by alterationsin hypochordal longitudinalis origin and insertion

For specific muscles at each phylogenetic level see caption for Figure 2.

LOCOMOTION IN RAY-FINNED FISHES

/Epuralt 1-3

93

Uroneural

Uroneural 2

Hypural 4

HypuralS

PUI + UI

Hypural IPU3

PU2

0-5 cm

FIG. 7. Caudal skeleton of Lepomis gibbosus. This morphology is typical of a generalized percomorph teleostcaudal skeleton (Gosline, 1961). Note the posteriorly directed (and expanded) hypurals, the three epurals,and the two small uroneural bones. The uroneurals are considerably smaller than those present in earlyteleosts (e.g., at Node 6 in Fig. 2). Note that the only true epaxial elements in this homocercal caudal fin arethe uroneurals and epurals; these support relatively few dorsal fin rays. The strain gauge results presentedin Figure 9 were recorded on hypural four. From Lauder (1982). Abbreviations: PUi_3, preural centra oneto three; U,, ural centrum one.

personal observations). Taxa that have atruly symmetrical internal caudal skeletonsuch as some scombroids (Fierstine andWalters, 1968) have lost the hypochordallongitudinalis muscle. Within the Teleos-tei, modifications of the caudal skeletonare common and include fusion of thehypural bones and reduction and fusion ofepurals and uroneurals. In most teleostfishes the hypurals have expanded and areoriented posteriorly, supporting fin raysboth above and below the precaudal ver-tebral axis (Fig. 7).

This procedure of mapping morpholog-ical novelties in the axial musculoskeletalsystem onto a phylogenetic hypothesis has

allowed the definition of a precise histor-ical sequence of structural transformation.The key general morphological results ofthis procedure are (1) that the hypochordallongitudinalis muscle differentiated priorto the origin of a homocercal tail in teleostfishes, (2) that the first intrinsic caudal mus-cles were associated with the ventral finrays, (3) that the hypochordal longitudi-nalis is derived from the dorsal portion ofthe flexor ventralis muscle, (4) that trueexternal tail symmetry (homocercy) inteleost fishes is a result of novel internalskeletal supports (epural and uroneuralbones) for dorsal fin rays, and (5) that mus-culature involved in expanding and con-

94 GEORGE V. LAUDER

Fltxordorsolis superior

Flexor dorsolis

Dorsalcartilage body

Interradialis

Supracarinalisposterior

Lateralissuperficialis

Flexorventralis externus

Haemal spineof P U 4

Haemal spineof PU3

Infracarinolisposterior

Ray I

Hypochordallongitudinalis

Flexor ventralis

Flexor ventralisinferior

Ventralcartilagebody

I cmRay I?

FIG. 8. Caudal musculature in Lepomis macrochirus. This lateral view of the middle and deep muscle layersshows the typical condition in the tail of a percomorph teleost with a homocercal tail. The ventral portionof the lateralis superficialis muscle has been removed to show the deeper musculature. The illustrated dorsalportion of this muscle passes under the hypochordal longitudinalis to attach to the heads of the dorsal finrays. Note the presence of both epaxial and hypaxial carinal muscles, and the hypochordal longitudinalismuscle. This muscle is the only intrinsic caudal muscle that connects ventral and dorsal components of thecaudal skeleton. From Lauder (1982).

trading the caudal fin rays (thus allowingmodulation of caudal area) evolved in twosteps, the first changes occurring in thedorsal region, with subsequent modifica-tions in the ventral aspect of the fin.

Functional hypotheses

The phylogenetic sequence of structuralmodification in the caudal musculoskeletalapparatus outlined above provides data onwhich to frame hypotheses of the evolutionof function. This is an important aspect ofthe interplay of historical and functionalanalyses: phylogenetic patterns suggestexplicit functional hypotheses. Proposedfunctional changes at successive hierarchi-cal levels on the cladogram of Figure 2 are

outlined in Table 1. Primitively within theActinopterygii, caudal fin ray movementwas coupled to myotomal contraction dueto the lack of distinct intrinsic caudal mus-cles. Caudal fin ray stiffness and positionwere presumably not greatly alterable dur-ing locomotion except as a direct conse-quence of myotomal contraction. At thelevel of the Ginglymodi (Fig. 2: Node 2),independent control of caudal fin raymotion was achieved with the origin of thelarge flexor ventralis muscle. This muscleallows lateral bending and control of finray motion independently of actions of thelateralis superficialis muscle.

At the next level (Fig. 2; Table 1: Node3) the first ability to modulate caudal fin

LOCOMOTION IN RAY-FINNED FISHES 95

area occurs. Only the epaxial pair of antag-onistic muscles (interradialis and supracari-nalis posterior) is present, indicating thatthe ventral fin rays were less capable ofadduction and abduction. Thus, althoughthe ability to control caudal fin ray move-ment in a medio-lateral direction is bestdeveloped for the ventral fin rays, the abil-ity to adduct and abduct the fin rays in thesagittal plane develops first in the dorsalfin rays.

At Nodes 4 and 5 (Fig. 2; Table 1) thedorsal fin rays receive their first epaxialskeletal supports, and the posteriorlydirected haemal spines (hypurals) areexpanded, marking the transition from theprimitive heterocercal fin to a homocercalshape. Finally, at Node 6, caudal fin areais alterable by movements of both dorsaland ventral rays as the hypaxial interra-dialis and carinal muscles are present. Atthis level, the hypochordal longitudinalisalso acquires an origin from the hypuralsand attachments to the first four dorsal finrays.

The historical sequence of hypochordallongitudinalis muscle origin and the dis-cordance between structural modificationsin the dorsal and ventral parts of the tailsuggests an important function for thehypochordal muscle in locomotion. Theexternally symmetrical teleost homocercaltail is internally asymmetrical and the dor-sal fin rays are not supported either by skel-etal elements or by flexor musculature tothe extent of the ventral fin rays. As thetail is swept from side to side, the dorsaledge will tend to trail the leading ventraledge. This should be especially true inclades possessing well-developed ventralflexors but weakly developed hypochordallongitudinalis muscles such as the gars(Ginglymodi). The phylogenetic develop-ment of the hypochordal longitudinalis iscorrelated with increasing epaxial supportfor the dorsal fin rays (Fig. 2; Table 1).This muscle may thus function to stiffenthe dorsal tail margin and generate a flat-ter tail surface that is not inclined to thehorizontal during locomotion. Under thishypothesis, the hypochordal longitudinalismuscle is essential for homocercal t^W func-tion as even the externally symmetrical

arrangement of caudal fin rays does not byitself counter internal structural asymme-try. A broader implication of this idea isthat teleost fishes should be able to mod-ulate their tail function by controlling thelevel of activity in the hypochordal muscle.If the hypochordal muscle is used duringlocomotion, then the tail should functionsymmetrically. If the hypochordal muscleis not activated, then the tail should func-tion asymmetrically and the dorsal marginis predicted to trail behind the ventral asthe caudal fin sweeps from side to side.This would tend to twist the caudal skel-eton so that the posterior edge of the dor-sal hypurals was bent to the side oppositeto that which the tail is being moved.

FUNCTIONAL MORPHOLOGY

Introduction

In order to provide quantitative exper-imental data bearing on the function of thehomocercal tail in ray-finned fishes and totest the above hypotheses, I conducted atwo-part experimental analysis designed toaddress the following questions. (1) Doesthe homocercal teleost tail actually func-tion symmetrically? (2) Does the teleost tailchange from symmetrical to asymmetricalfunction depending on locomotor mode?(3) Is the hypochordal longitudinalis mus-cle used during locomotion and is activityin this muscle correlated with symmetricaltail function? By symmetrical tail function,I mean that the caudal fin is generatingforces oriented anteroposteriorly and thatthe dorsal lobe of the tail is functioning inconcert with the ventral and not trailingbehind or generating less propulsive force.

The first set of experiments was designedto address questions (1) and (2) above. Ifone could directly measure bending inhypural bones, then an indication of howthe caudal fin was being used during dif-ferent locomotor behaviors could beobtained. Such measurements were madeby implanting a strain gauge onto hypuralnumber 4 and directly assessing the pat-tern of hypural deformation during bothcontinuous locomotion and fast starts.Measurement of caudal muscle functionwas accomplished by electromyographicrecordings of both myotomal and intrinsic

96 GEORGE V. LAUDER

muscles (such as the hypochordal longitu-dinalis) during both continuous locomo-tion and fast starts.

Both of these approaches have beenapplied to the caudal fin of sunfishes (fam-ily Centrarchidae). While this group ofteleost fishes is phylogenetically derivedwithin the Teleostei, these fishes constitutean excellent system within which to gatherexperimental data and test initial explan-atory hypotheses of caudal function. Agreat deal of continued research is clearlyneeded to extend these experimentalresults to other ray-finned clades.

Strain gauge analysisPatterns of hypural deformation during

locomotion were studied in Lepomis gibbosus(Lauder, 1982). The caudal skeleton (Fig.7) has five posteriorly oriented hypuralbones with hypurals one and four beingthe largest. Two uroneurals and threeepurals form the epaxial components ofthe skeleton (Fig. 7). In lateral view, themusculature of the caudal fin (Fig. 8) istypical of the teleost homocercal condition(Nursall, 1963; Winterbottom, 1974; Lau-der, 1982). A large lateralis superficialand its aponeurosis cover the intrinsic mus-cles laterally (the ventral portion of the lat-eralis and the aponeurosis have beenremoved in Fig. 8) and attach to the headsof the fin rays. Infra- and supracarinalisposterior muscles attach to cartilage ele-ments anterior to the ventral and dorsalprocurrent rays. The hypochordal longi-tudinalis muscle originates mainly fromhypural one and narrows to four tendonsthat insert on dorsal fin rays one to four(Fig. 8).

Rosette strain gauges were bonded toone side of hypural four (Fig. 7) in fourspecimens (see Lauder [1982] for details ofthe experimental procedures used) in orderto measure the pattern of bone deforma-tion during locomotion. Rosette straingauges measure bone deformation alongthree axes simultaneously and thus allowthe calculation of the principal strain angleand an estimate of the loading situation.Fishes bearing the strain gauges wereallowed to swim freely in an aquarium forrecordings of bone strain during continu-

ous locomotion, and were startled intorapid accelerations for fast-start record-ings.

Figure 9 shows the results of the strainrecordings on hypural four. During fast-start accelerations, the principal strains areoriented parallel and perpendicular to thebody axis (Fig. 9: FS). These data are con-sistent with a loading pattern of lateralbending with no twisting of hypural fourabout its attachment to the caudal verte-brae. This pattern is the expected one ifthe caudal fin is functioning symmetrically.The bone strain found during continuouslocomotion is quite different from the fast-start pattern, however (Fig. 9: CLC, CLI).The axes of compression and tension as thetail is swept from side to side are inclinedat an angle of 35° to the horizontal. Thisindicates that the homocercal tail is func-tioning asymmetrically and that the dorsalhypurals and tail lobe are contributing lessforce to propulsion than the ventral lobe.The observed deformation pattern wouldbe produced if the posterodorsal corner ofhypural four (the site of fin ray attachment)were lagging behind the rest of the caudalskeleton as it was moving laterally. Theangles of deformation of the hypural wereidentical for both tail motion to the rightand left (Lauder, 1982) indicating thatstrain gauge implantation was not affectingthe strain pattern.

Questions (1) and (2) raised at the startof this section can now be answered. Theseresults clearly indicate that the homocercalteleost tail does function asymmetricallyduring slow steady swimming and that tailfunction does change with locomotor mode(fast starts). Internal skeletal asymmetryappears to have a major role in caudal func-tion during normal steady swimming. It isnot necessarily true that a homocercal tailfunctions symmetrically, and one must nowquestion proposals of increased caudal effi-ciency in homocercal tails during contin-uous locomotion.

What is causing the shift in caudal func-tion between fast starts and continuouslocomotion? The hypochordal longitudi-nalis muscle is ideally placed to cause sucha change in strain pattern. The line ofaction of the hypochordal muscle crosses

LOCOMOTION IN RAY-FINNED FISHES 97

CLC

Anterior — — — — — — — S— — — — ^ — — — — — — — Posterior

CLC

FIG. 9. Orientation of the strain patterns on hypural four recorded during both continuous locomotion (CL)and fast-start accelerations (FS). The direction of the arrowheads indicates tensile or compressive strains, andthe relative length of the perpendiculars reflects the ratio of compressive to tensile strain on hypural four.Note that during fast starts the axes of deformation are oriented parallel and perpendicular to the body axis,while during continuous locomotion the strain axes are inclined at about a 30" angle to the horizontal. FromLauder (1982). Abbreviations: CLC, continuous locomotion, data from contralateral tail strokes (movementsof the tail to the side on which the strain gauge was not attached); CLI, ipsilateral data during continuouslocomotion.

the posterior edge of hypural four to attachto the four dorsal fin rays. Activity of thismuscle during locomotion would bend theposterodorsal edge of hypural four antero-ventrally and thus rotate the obliquelyinclined strain shown in Figure 9 dorsallytowards the horizontal. This hypothesispredicts that the major difference betweensymmetrical function during fast starts andasymmetrical tail function during contin-uous locomotion is due to differential activ-ity in the hypochordal longitudinalis mus-cle.

Electromyographic analysis

This prediction (and question (3) raisedat the start of this section) was tested byrecording muscle electrical activity duringlocomotion in the bass (Micropterus sal-moides; three specimens) and the pumpkin-seed sunfish (Lepomis gibbosus; two speci-mens). Muscles from which activity patterns

were obtained include deep portions of themidbody myotomes, myotomal segmentsin the caudal peduncle, and the hypochor-dal longitudinalis and flexor dorsalisintrinsic caudal muscles. In order to pro-vide an indication of tail movement, a smallunipolar electrode was implanted subcu-taneously in the posterodorsal aspect of thecaudal peduncle at the point where thedorsal fin rays attach to the epural bones.An impedance converter (see Lauder andShaffer [1985] for details) was used to passa high-frequency signal between this elec-trode and a metal plate on the side of theflow tank. As the tail moved from side toside, the impedance between the tail elec-trode and the metal plate changed, pro-viding a direct measure of tail oscillation.Recordings were made as fishes swam in aflow tank (28 by 18 cm working section,designed following the plans in Vogel andLaBarbera [1978]). At slow speeds, bass

98 GEORGE V. LAUDER

and pumpkinseed swam continuously usingbody and caudal fin propulsion. At highflow velocities (above 0.8 m/sec) fishes wereunable to maintain their position andresorted to a burst and glide locomotorpattern. These bursts involved rapid accel-erations and were used to assess caudalmuscle function during near maximalexertion. Fishes could use the burst andglide pattern only for 10-30 sec beforebecoming exhausted.

Figure 10 shows representative data fromfour muscles during locomotion in the bass.As expected from the strain gauge record-ings, the hypochordal longitudinalis hasonly an extremely low level of activity, evenduring high-speed continuous locomotion.However, during rapid accelerations, thehypochordal longitudinalis is stronglyactive (Fig. 10: HL) and is repeatedly usedas the bass attempts to maintain its positionin the flow.

DISCUSSION AND PROSPECTUS

The historical and functional analysespresented here provide a basis for reinter-preting the evolution of the caudal axialmusculoskeletal system in ray-finned fishes,and suggest many further lines of inquiry.The experimental data make clear that theview that the homocercal teleost tail func-tions symmetrically must be questioned.The pattern of homocercal tail function inteleost fishes (at least those studied exper-imentally) strongly depends on the pres-ence of electrical activity in the hypochor-dal longitudinalis muscle. The intrinsicmuscles of the tail cannot be ignored asthey have in past discussions of caudal finfunction and evolution.

The historical, phylogenetic analysisindicates that the hypochordal longitudi-nalis muscle originated as a division of theflexor ventralis (a hypaxial muscle) and thatthe ability to alter the area of the caudalfin was achieved first for the dorsal fin rays

(Fig. 2: Node 3; Fig. 6). Only later did thecapability of altering the area of the ventralportion of the caudal fin evolve. Further-more, there is a clear historical relation-ship between the origin of internal epaxialsupports and modifications of the hypo-chordal longitudinalis muscle origin andinsertion.

A synthesis of both the phylogenetic andfunctional approaches to the axial muscu-loskeletal system of ray-finned fishes sug-gests a number of avenues for furtherresearch. A much broader base of com-parative experimental data is necessary tofurther test the generality of the hypoth-esis that the hypochordal longitudinalismuscle plays a major role in determiningthe functional properties of the caudal fin.Marshall (1971, pp. 30-31) has also sug-gested that the hypochordal longitudinalismuscle is integral to caudal fin kinematics,but his hypotheses have yet to be tested.Quantitative electromyographic data areneeded from the intrinsic caudal andmyotomal musculature in Polypterus, Lepi-sosteus, and Amia as well as for primitiveteleosts such as osteoglossomorphs. Datasuch as these would permit functional con-sequences of structural features to bemapped as characters onto the phyloge-netic hypothesis (Fig. 2) much more pre-cisely than can now be done. It is clear thatone cannot explain the complexity ofintrinsic caudal musculature in teleost fisheson the basis only of steady swimming andburst and glide locomotion. Considerableresearch will need to be done on maneu-verability and use of the caudal fin indynamic stability before the full signifi-cance of intrinsic muscle complexity willbecome apparent.

The exact role of the hypochordal lon-gitudinalis muscle in mediating tail func-tion needs detailed investigation, both byprecise correlative studies of tail kinemat-ics in relation to hypochordal longitudi-

FIG. 10. Representative electromyograms from locomotion in the bass (Micropterus salmoides). Upper panel:recordings from fast continuous swimming. Lower panel: recordings from a rapid acceleration (burst). Notethe increase in amplitude of the hypochordal longitudinalis muscle during fast starts. Vertical scale is in fiV'102. Abbreviations: APM, anterior peduncular m)otome; HL, hypochordal longitudinalis muscle; IMPED,impedance trace of tail position (see text); PPM, posterior peduncular myotome.

LOCOMOTION IN RAY-FINNED FISHES 99

APM tl r

PPM

- l

HL

IMPED o

100 ms

B

APM o

PPM 0

HL o

IMPED 0

4 Contralateral

Ipsilateral

100 ms

100 GEORGE V. LAUDER

nalis muscle activity, and by experimentalmanipulations. For example, the tendonsof the hypochordal longitudinalis could besevered at their attachments to the dorsalfin rays, and the effect on caudal functionobserved. If the view presented here is cor-rect, eliminating the hypochordal longi-tudinalis should reduce fast-start perfor-mance without substantially modifying theefficiency of continuous locomotion. Elim-inating the hypochordal longitudinalis ispredicted to cause the dorsal fin rays totrail the ventral rays during fast starts andproduce an asymmetrical thrust duringrapid acceleration. Aleyev (1969, 1977),Bainbridge (1963), and Marshall (1971)have all noted that the dorsal and ventralfin rays in the teleost tail may exhibit dif-ferent kinematic patterns, but the conse-quences of these observations have neverbeen pursued, especially with respect to thepossibility that hypochordalis muscle mightbe causally implicated in caudal fin kine-matics.

The results presented here raise thequestion: exactly what is a homocercal tail?Since the sunfish tail, exhibiting the clas-sical homocercal shape, has been shown tofunction asymmetrically, then the defini-tion of a homocercal tail should only applyto external shape, referring specifically toequal dorsal and ventral fin ray lobes in thetail. Internally, there is considerable struc-tural asymmetry, and neither dorsal/ven-tral nor epaxial/hypaxial components aremirror images of each other. There isnothing about the external shape of thecaudal fin alone that enables us to predictits functional properties. Function andlocomotor performance will be a conse-quence of the interaction of external shape,internal skeletal morphology, and thedesign and relative activity of intrinsic cau-dal musculature.

Any morphological system has a historythat must be taken into account if we areto fully understand the relationshipbetween form and function. Withoutunderstanding the historical patterns thathave led to current structures, we risk mis-taking correlations between present-daystructures, environments, and functions fora causal relationship. Determining causal

interrelationships of form and functionrequires an historical analysis (Lauder,1989). The axial musculoskeletal system ofray-finned fishes is no exception: func-tional and phylogenetic approaches aremutually illuminating. Yet such a com-bined approach is still in its infancy, and amore general understanding of vertebrateaxial musculoskeletal systems awaits thesynthesis of historical and functional infor-mation.

ACKNOWLEDGMENTS

I thank Peter Wainwright and SteveReilly for comments on the manuscript,and Ellengene Peterson for her efforts inorganizing the symposium. Peter Wain-wright provided crucial assistance with theelectromyographic experiments. Specialthanks to Joseph Biela, Ed Lau, and RudyLimburg who built the flow tank and mod-ified its design on short notice. Figures 3—6 were drawn by Clara Richardson. Thisresearch was completed during the tenureof grant NSF BSR 84-20711 (SystematicBiology Program), which supported thisattempt to integrate historical and exper-imental analyses.

REFERENCES

Affleck, R. J. 1950. Some points in the function,development, and evolution of the tail in fishes.Proc. Zool. Soc. London 120:349-368.

Alexander, R. McN. 1967. Functional design in fishes.Hutchinson, London.

Aleyev, Yu G. 1969. Function and gross morphology infish. Keter Press, Jerusalem.

Aleyev, Yu G. 1977. Nekton. Junk, The Hague.Bainbridge, R. 1963. Caudal fin and body movement

in the propulsion of some fish. J. Exp. Biol. 40:23-56.

Blake, R. W. 1983. Fish locomotion. Cambridge Uni-versity Press, Cambridge.

Bone, Q. 1966. On the function of the two types ofmyotomal muscle fibre in elasmobranch fish. J.Mar. Biol. Ass. U.K. 46:321-349.

Bone, Q. 1978. Locomotor muscle. In W. S. Hoarand D. J. Randall (eds.), Fish physiology, Vol. 7,pp. 361-423. Academic Press, New York.

Bone, Q. 1989. Evolutionary patterns of axial musclesystems in some invertebrates and fish. Amer.Zool. 29:5-18.

Breder, C. M. 1926. The locomotion of fishes. Zoo-logica, N.Y. 4:159-256.

Cowan, G. I. 1969. The cephalic and caudal mus-culature of the sculpin Myoxocephalus polyacantho-cephalus (Pisces: Cottidae). Can. J. Zool. 47:841-850.

LOCOMOTION IN RAY-FINNED FISHES 101

Fierstine, H. L. and V. Walters. 1968. Studies inlocomotion and anatomy of scombroid fishes.Mem. South. Calif. Acad. Sci. 6:1-31.

Gardiner, B. G. 1963. Certain palaeoniscoid fishesand the evolution of the actinopterygian snout.Bull. Br. Mus. Nat. Hist. (Geol.) 8:255-325.

Gardiner, B. G. 1984. The relationships of thepalaeoniscid fishes, a review based on new spec-imens of Mimia and Moythomasia from the UpperDevonian of Western Australia. Bull. Br. Mus.Nat. Hist. (Geol.) 37:173-428.

Goodrich, E. S. 1930. Studies on the structure and devel-opment of vertebrates. Macmillan, London.

Gosline,W. A. 1961. The perciform caudal skeleton.Copeia 1961:265-270.

Gosline, W. A. 1971. Functional morphology and clas-sification of teleostean fishes. University of HawaiiPress, Honolulu.

Gray,J. 1933a. Studies in animal locomotion I. Themovement of fish with special reference to theeel.J. Exp. Biol. 10:88-104.

Gray,J. 19336. Studies in animal locomotion III. Thepropulsive mechanism of the whiting {Gadus mer-langus).]. Exp. Biol. 10:391-400.

Gray.J. 1968. Animal locomotion. W. W. Norton, NewYork.

Greenwood, P. H. 1967. The caudal fin skeleton inosteoglossid fishes. Ann. Mag. Nat. Hist. 9:581-597.

Johnston, I. A. 1980. Contractile properties offishfast muscle fibers. Mar. Biol. Lett. 1:323-328.

Jollie, M. 1972. Chordate morphology. R. Krieger, NewYork.

Lauder, G. V. 1982. Structure and function in thetail of the pumpkinseed sunfish (Lepomisgibbosus).J. Zool. (London) 197:483-495.

Lauder, G. V. 1989. Biomechanics and evolution:Integrating physical and historical biology in thestudy of complex systems. In J. A. Rayner (ed.),Biomechanics in evolution, Chapter 1. CambridgeUniv. Press, Cambridge. (In press)

Lauder, G. V. and K. F. Liem. 1983. The evolutionand interrelationships of the actinopterygianfishes. Bull. Mus. Comp. Zool. 150:95-197.

Lauder, G. V. and H. B. Shaffer. 1985. Functionalmorphology of the feeding mechanism in aquaticambystomatid salamanders. J. Morph. 185:297-326.

Liem, K. F. 1970. Comparative functional anatomyof the Nandidae (Pisces, Teleostei). Fieldiana Zool.56:1-166.

Lighthill.J. 1971. Large-amplitude elongated-bodytheory of fish locomotion. Proc. Roy. Soc. Lon-don B 179:125-138.

Lindsey.C. C. 1978. Form, function, and locomotoryhabits in fish. In W. Hoar and D. J. Randall (eds.),Fish physiology, Vol. 7, pp. 1-100. Academic Press,New York.

Lund, R. 1967. An analysis of the propulsive mech-anisms of fishes, with references to some fossilactinopterygians. Ann. Carneg. Mus. 39:195-218.

Marshall, N. B. 1971. Explorations in the life of fishes.Harvard Univ. Press, Cambridge.

Nursall.J. R. 1963. The caudal musculature of Hoplo-

pagrus guntheri Gill (Perciformes: Lutjanidae).Can.J. Zool. 41:865-880.

Nybelin, O. 1973. Comments on the caudal skeletonof actinopterygians. In P. H. Greenwood, R. S.Miles, and C. Patterson (eds.), Interrelationships offishes, pp. 369-372. Academic Press, London.

Olson, E. C. 1971. Vertebrate paleozoology. John Wiley,New York.

Patterson, C. 1968. The caudal skeleton in lowerliassic pholidophorid fishes. Bull. Br. Mus. Nat.Hist. Geol. 16:203-239.

Patterson, C. 1973. Interrelationships of holosteans.In P. H. Greenwood, R. S. Miles, and C. Patterson(eds.), Interrelationships of fishes, pp. 233-305.Academic Press, London.

Patterson, C. 1977. The contribution of paleontol-ogy to teleostean phylogeny. In M. K. Hecht, P.C. Goody, and A. D. Walker (eds.), Major patternshi vertebrate evolution, pp. 579-643. Plenum, NewYork.

Patterson, C. 1982. Morphology and interrelation-ships of primitive actinopterygian fishes. Amer.Zool. 22:241-259.

Patterson, C. and D. E. Rosen. 1977. Review of ich-thyodectiform and other Mesozoic teleost fishesand the theory and practice of classifying fossils.Bull. Amer. Mus. Nat. Hist. 158:81-172.

Pearson, D. M. 1982. Primitive bony fishes, with spe-cial reference to Cheirolepis and palaeonisciformactinopterygians. Zool. J. Linn. Soc. London 74:35-67.

Romer, A. S. and T. S. Parsons. 1986. The vertebratebody. Saunders, New York.

Schaeffer, B. and N. G. McDonald. 1978. Redfieldiidfishes from the Triassic-Liassic Newark super-group of eastern North America. Bull. Amer.Mus. Nat. Hist. 159:131-173.

Schultze, H.-P. and G. Arratia. 1986. Reevaluationof the caudal skeleton of actinopterygian fishes:I. Lepisosteus and Amia. J. Morph. 190:215-241.

Videler.J. and F. Hess. 1984. Fast continuous swim-ming of two pelagic predators, saithe (Pollachiusvirens) and mackerel {Scomber scombrus): A kine-matic analysis. J. Exp. Biol. 109:209-228.

Vogel, S.and M. LaBarbera. 1978. Simple flow tanksfor research and teaching. Bioscience 28:638-643.

Wainwright, S. A. 1983. To bend a fish. In P. W.Webb and D. Weihs (eds.), Fish biomechanics, pp.68-91. Praeger, New York.

Webb, P. W. 1975. Hydrodynamics and energeticsoffish propulsion. Bull. Fish. Res. Bd. Can. 190:1-158.

Webb, P .W.I 978a. Fast-start performance and bodyform in seven species of teleost fish. J. Exp. Biol.74:211-226.

Webb, P. W. 19786. Hydrodynamics: Nonscombroidfish. In W. Hoar and D. J. Randall (eds.), Fishphysiology, Vol. 7, pp. 189-237. Academic Press,New York.

Weihs, D. 1973. The mechanism of rapid starting ofslender fish. Biorheolology 10:343-350.

Weihs, D. 1989. Design features and mechanics of

102 GEORGE V. LAUDER

axial locomotion in fish. Amer. Zool. 29:151-160.

Whitehouse, R. H. 1910. The caudal fin of theTeleostomi. Proc. Zool. Soc. London 1910:590-626.

Wiley, E. O. 1976. Phylogeny and biogeography offossil and Recent gars (Actinopterygii: Lepisos-

teidae). Univ. Kansas Mus. Nat. Hist. Misc. Publ.64:1-111.

Winterbottom, R. 1974. A descriptive synonymy ofthe striated muscles of the Teleostei. Proc. Acad.Nat. Sci. Phil. 125:225-317.

Wu, T., C. J. Brokaw, and C. Brennen. 1975. Swim-ming and flying in nature. Plenum, New York.