Patterns of Evolution in the Feeding Mechanism of Actinopterygian

AMER. ZOOL., 22:275-285 (1982)

Patterns of Evolution in the Feeding Mechanism ofActinopterygian Fishes1

GEORGE V. LAUDER

Department of Anatomy, University of Chicago, Chicago, Illinois 60637

SYNOPSIS. Structural and functional patterns in the evolution of the actinopterygianfeeding mechanism are discussed in the context of the major monophyletic lineages ofray-finned fishes. A tripartite adductor mandibulae contained in a maxillary-palatoquad-rate chamber and a single mechanism of mandibular depression mediated by the obliquusinferioris, sternohyoideus, and hyoid apparatus are primitive features of the Actinopte-rygii. Halecostome fishes are characterized by having an additional mechanism of man-dibular depression, the levator operculi—opercular series coupling, and a maxilla whichswings anteriorly during prey capture. These innovations provide the basis for feedingby inertial suction which is the dominant mode of prey capture throughout the haleco-stome radiation. A remarkably consistent kinematic profile occurs in all suction-feedinghalecostomes. Teleost fishes possess a number of specializations in the front jaws includinga geniohyoideus muscle, loss of the primitive suborbital adductor component, and a mobilepremaxilla. Structural innovations in teleost pharyngeal jaws include fusion of the dermaltooth plates with endoskeletal gill arch elements, the occurrence of a pharyngeal retractormuscle, and a shift in the origin of the pharyngohyoideus. These specializations relate toincreased functional versatility of the pharyngeal jaw apparatus as demonstrated by anelectromyographic study of pharyngeal muscle activity in Esox and Ambloplites. The majorfeature of the evolution of the actinopterygian feeding mechanism is the increase instructural complexity in both the pharyngeal and front jaws. Structural diversification isa function of the number of independent biomechanical pathways governing movement.

INTRODUCTION

The evolution of the feeding mechanismin ray-finned fishes (Actinopterygii) pro-vides perhaps the best documented ex-ample in the Vertebrata of change in astructurally and functionally complex sys-tem. In the twenty years since the last re-view of the evolution of the feeding mech-anism in ray-finned fishes (Schaeffer andRosen, 1961), knowledge of both the his-torical pattern of diversification and therelation between structure and functionhas increased tremendously. Phylogeneticanalyses of actinopterygian evolutionarypatterns have provided an excellent base-line of information on the historical se-quence of structural change (Greenwoodet al., 1966, 1973; Patterson, 1977, 1982;Patterson and Rosen, 1977; Rosen, 1982),and as the discipline of experimental func-tional morphology has developed, a cor-responding increase has occurred in theanalysis of the relationship between form

1 From the Symposium on Evolutionary Morphologyof the Actinopterygian Fishes presented at the AnnualMeeting of the American Society of Zoologists, 27-30 December 1980, at Seattle, Washington.

and function in fishes and in the proposaland testing of functional explanations forstructure (Alexander, 1966, 1967, 1970;Anker, 1974; Lauder, 1979; Liem, 1970;Osse, 1969). Functional analysis has be-come increasingly sophisticated and tech-niques such as high-speed cinematography(Elshoud-Oldenhave and Osse, 1976; Gro-becker and Pietsch, 1979; Nyberg, 1971),electromyography (Ballintijn et al., 1972;Lauder, 1980a; Liem, 1973; Liem andOsse, 1975; Vandewalle, 1979), straingauges (Lauder and Lanyon, 1980), andpressure transducers (Alexander, 1970;Lauder, 19806, c; Osse and Muller, 1981)have largely obviated the need to basefunctional considerations on manipula-tions of preserved or freshly dead speci-mens. As a result, many hypotheses aboutthe functional significance of morpholog-ical features in the actinopterygian skullhave been tested, and many previously un-suspected relationships have emerged.

In this paper, I will focus on structuraland functional specializations in the evo-lution of the actinopterygian feedingmechanism as they are reflected in nestedsets of monophyletic lineages. I will em-

275

276 GEORGE V. LAUDER

FIG. 1. Structural network in the head of a primitiveactinopterygian (A), a primitive halecostome (B), anda percomorph (C) to show the biomechanical path-ways governing mouth opening, suction feeding, andjaw protrusion functions. Homologous biomechanicalpathways are similarly numbered. Note the increasein complexity of the structural network in actinop-terygian evolution. Only the function of jaw protru-sion is shown in (C); the primitive functions of mouthopening and suction feeding are omitted for clarity.Solid rectangles = bony elements; dashed rectan-gles = ligaments; parallelograms = muscles. Arrowsrun from the muscle to the bone of insertion; double-headed arrows indicate ligamentous connections be-tween bony elements. Three dimensional rectanglesindicate major functions which are realized (r, ar-rows) by the biomechanical couplings indicated. Thisfigure is not the same as the diagrams depicting thepattern of interrelationships and functional influ-ences (see Dullemeijer, 1974, Fig. 62). Abbreviations:AMI, division Al of the adductor mandibulae: A OP,

MX

BM

HVOID

FIG. 2. Reconstruction of the superficial lateral (A)and ventral (B) cranial musculature in the paleonis-ciform fish Moythomasia nilida Gross. Osteological ele-ments modified after Jessen (1968). The anteriorbranchiostegal rays have been removed in the ventralview to show the reconstructed throat musculature,and the maxilla in the lateral view has been partiallyremoved to reveal the adductor musculature. Thepaired sternohyoideus muscles lie deep to the inter-mandibularis posterior and interhyoideus and are notvisible in this view. This reconstruction results fromdeducing the most parsimonious primitive arrange-ment of adductor muscle character states in livingactinopterygians. Abbreviations: AMa, anterior (sub-orbital) division of the adductor mandibulae; AMm,medial adductor division; AMp, posterolateral ad-ductor division; BM, branchiomandibularis muscle;CL, cleithrum; CLAV, clavicle; EP, epaxialis; IH,interhyoideus; IMp, intermandibularis posteriormuscle; IO, infraorbital bone; MD, mandible; MX,maxilla; OBI, obliquus inferioris; OBS, obliquussuperioris; OP, operculum; POP, preoperculum.

adductor operculi muscle; EM, epaxial muscles; HY,hypaxial (obliquus inferioris) musculature; IHL, in-teroperculohyoid ligament; IML, interoperculoman-dibular ligament; LAP, levator arcus palatini muscle;LOP, levator operculi muscle; MHL, mandibulohyoidligament; SH. sternohvoideus muscle.

FEEDING MECHANISMS IN RAY-FINNED FISHES 277

phasize characteristic features of the Ac-tinopterygii, Halecostomi, Teleostei, Neo-teleostei, and Percomorpha with the goalof suggesting certain general propositionsabout the nature of change in structuraland functional networks within lineages(also see Lauder, 1981).

PRIMITIVE FEATURES OF THEACTINOPTERYGIAN FEEDING MECHANISM

Mouth opening in primitive actinopte-rygians is mediated by two musculoskeletalcouplings (Figs. 1A, 2): the epaxial mus-cles—neurocranium coupling which ele-vates the head (Fig. 1A: coupling 2), anda ventral coupling involving the hypaxialmusculature, cleithrum, sternohyoideus,and hyoid apparatus (Fig. 1A: coupling 1)which causes mandibular depression.Depression of the lower jaw is effected byretraction of the hyoid apparatus (by thesternohyoideus and obliquus inferiorismuscles) which exerts a posterodorsalforce on the mandible via the mandibulo-hyoid ligament (Fig. 1A: MHL). This pos-terodorsal force is applied at the insertionof the mandibulohyoid ligament ventral tothe quadratomandibular articulation andthus causes mandibular depression (alsosee Lauder, 1980a, Fig. 18; 1980d)- Thismechanism of mandibular depression isalso found in lungfishes, coelacanths, andsharks and is thus primitive for the Tel-eostomi.

A reconstruction of the jaw musculaturein a palaeoniscoid is illustrated in Figure2. Laterally, the adductor mandibulae isdivided into three divisions: an anterior(suborbital) division, a medial division, anda posterolateral division (Fig. 2A). Thesethree adductor components are hypothe-sized to be homologous with similarly lo-cated muscle divisions in Polypterus, Lepi-sosteus, and Amia (see Lauder [1980a, e] fora more extensive discussion of muscle ho-mologies). An intramandibular adductor(Aw) occupies the mandibular adductorfossa. Ventrally, the paired sternohyoideusmuscles extend anteriorly to insert on theurohyal. A flat, wide intermandibularisposterior spans the mandibular rami (Fig.2B: IMp) and the interhyoideus extendsanteriorly from the ceratohyal and dorsal

surface of the branchiostegal rays to insertin the fascia dorsal to the intermandibu-laris posterior. The hyohyoideus muscu-lature of halecostomes appears to be de-rived from the interhyoideus muscle fibersof primitive actinopterygians (Fig. 2B:IH).

The skull of primitive actinopterygianspossesses only a few mobile elements (Blot,1978; Saint-Seine, 1956). The maxilla andpremaxilla are firmly attached to the otherdermal skull bones and the opercle, sub-opercle, and branchiostegal rays have lim-ited lateral mobility. The oblique angle ofthe suspensory apparatus (reflected by theposition of the preoperculum, Fig. 2A:POP), results in a distinctly postorbital jawarticulation and limited lateral expansion.The three adductor mandibulae divisionsare contained in a postorbital maxillary-palatoquadrate chamber.

The experimental study of prey capturein the primitive living actinopterygians Po-lypterus and Lepisosteus (Lauder, 1980a;Lauder and Norton, 1980) has revealedthe importance of synchronous activity inthe obliquus inferioris and sternohyoideusmuscles for mouth opening. The ventraldivision of the hypaxialis (=obliquus infer-ioris) stabilizes the pectoral girdle so thatthe primary effect of the sternohyoideusis to cause posteroventral hyoid rotation,thus opening the mouth. Experimentalanalysis reveals no evidence for the pos-terior movement of the pectoral girdlewhich has been suggested to be involvedin feeding (Hutchinson, 1973; Schaefferand Rosen, 1961; Tchernavin, 1953). Theventral throat musculature plays little rolein mediating mouth opening. The inter-mandibularis posterior, interhyoideus, andbranchiomandibularis (Fig. 2B) are pri-marily active during chewing and intraoralmanipulation of prey items, and may beused to control fluid flow through the oralcavity. These ventral muscles are not in-volved in opening the mouth.

Primitive actinopterygian fishes (e.g.,Cheirolepis, Moythomasia) possessed a cephalicmusculoskeletal system which is consider-ably less complex mechanically than thatof the Halecostomi. There were relativelyfew mobile elements in the skull and ex-


pansion of mouth cavity volume duringfeeding was probably quite small due tothe limited lateral mobility of the suspen-sorium, opercle, and branchiostegal ele-ments. Water flow through the mouth cav-ity during feeding may thus have beenprimarily controlled by body velocity, andnot by movement of skull elements.

THE HALECOSTOME FEEDING MECHANISM

The Halecostomi share two major struc-tural innovations related to the feedingmechanism and possess a network of struc-tural connections in the head which is con-siderably more complex than that of thepaleoniscoid fishes (Fig. IB). Halecostomefishes share two independent biomechan-ical pathways mediating lower jaw depres-sion: the primitive coupling involving thehypaxial musculature, pectoral girdle,sternohyoideus, and hyoid apparatus (Fig.IB: coupling 1), and a second, new cou-pling involving the opercular apparatus(Fig. IB: coupling 2). This opercular seriescoupling is retained in nearly all of the25,000 species of halecostomes and is a re-markably persistent component of thestructural network of the head.

Mandibular depression by levation ofthe operculum is accomplished by contrac-tion of the levator operculi muscle whichis derived from the adductor operculi ofprimitive actinopterygians (Fig. 1A, B:AOP, LOP). The levator operculi causes adorsal rotation of the opercular series(opercle, subopercle, and interopercle)which is applied as a posterodorsal forceon the mandible via the interoperculoman-dibular ligament (Fig. IB: IML). Both thelevator operculi and interoperculum rep-resent structural specializations at the hal-ecostome level. The consequence of hav-ing two biomechanically independentpathways mediating mandibular depres-sion is a dissociation of the primitive hyoidcoupling (Fig. 1A, B: coupling 1) fromobligatory mouth opening functions. Thisallows changes in the timing of hyoiddepression in relation to mouth openingand increases the versatility of control offluid movement through the oral cavity.

Primitive halecostomes possess two oth-er structural innovations which relate to

control of fluid flow and suction feeding.The maxilla, primitively firmly attached tothe neurocranium and forming the lateralwall of the adductor chamber (Fig. 2A;Gardiner, 1963; Schaeffer and Rosen,1961) is free from the cheek and pivots ona medially directed process posterior to thevomer. High-speed cinematography ofprey capture by Amia calva and severalprimitive teleosts (Lauder, 1979) has re-vealed that the maxilla swings anteriorlyon its neurocranial pivot as the mouthopens, and by preventing fluid inflowthrough the corners of the mouth, resultsin anteroposteriorly oriented streamlineswhich increase the velocity of water move-ment from in front of the mouth into theoral cavity (see Fig. 3A: MX). Secondly, anincrease in the volume change within theorobranchial cavity results both from thevertical orientation of the suspensoriumand from increased ventral mouth cavityexpansion as a result of a greater range ofhyoid depression. A mobile maxilla, poten-tially large orobranchial volume changes,and increased kinematic versatility, arefeatures which are maintained in suctionfeeding fishes throughout the halecostomeradiation.

The kinematic pattern which is charac-teristic of primitive halecostome fishes isalso maintained in all generalized preda-ceous fishes studied to date (Figs. 3, 4). Asthe mouth begins to open during feeding,the operculum and branchiostegal rays areadducted against the pectoral girdle (Fig.3B: frames 1 and 2), preventing water in-flow. Maxillary swing and opercular leva-tion reach a peak nearly synchronouslywith peak mouth opening which is fol-lowed by hyoid depression and operculardilation. The mouth then rapidly closeswhile hyoid depression, suspensorial ab-duction, and opercular dilation return totheir initial positions (Figs. 3, 4). The pro-cess of buccal compression thus involves asequence of movement different frommouth cavity expansion.

Patterson (1973) considers the extinctgroups Parasemionotidae and Semionoti-dae to represent primitive grades of hale-costome organization. Both groups in-clude forms which show the structural


OPERCULAR DILATION (cm)

LOWER JAW ANGLE (degrees)

ANGLE OF MAXILLA (degrees)

1 3 5 7 9 11 13 15 17 19 21 23

FRAME NUMBER (TIME) ^ ^

FIG. 3. Prey capture in Atnia calva as seen in lateral(A) and ventral (B) views. This figure is traced fromframes of high-speed films of two separate prey cap-ture events. (Modified from figures 13 and 14 of Lau-der [1980a].) Note the delay in opercular and bran-chiostegal expansion until the mouth is nearlycompletely open, and the anteropostenor sequenceof peak excursion in mouth opening, hyoid depres-sion, and opercular dilation. Abbrevations: GP, gularplate; MX, maxilla.

correlates of a suction feeding mechanism.For example, Lepidotes has a well-devel-oped interopercle and a free maxilla bear-ing a prominent medial process (personalobservation on MCZ 5304, Lepidotes elven-sis). The hyoid apparatus of Lepidotes alsoappears to be very similar to that of otherhalecostomes.

THE TELEOST FEEDING MECHANISM

Major structural features

The teleost feeding mechanism is distin-guished from that of more primitive hal-ecostomes by the division of the premaxillainto a mobile lateral toothed portion and

FIG. 4. Pattern of jaw bone movement in Salvelmusfontinalts to illustrate the primitive halecostome ki-nematic profile. The relative sequence of bone move-ment at the strike is very similar in all predaceoushalecostomes that have been studied experimentally(see text).

a medial portion which becomes associatedwith the ethmoid complex (Patterson,1973). In a number of predaceous teleosts{e.g., Hoplias, Salmo) the premaxilla has be-come secondarily firmly attached to theneurocranium, but the primitive conditionfor teleosts as exemplified by Pholidopho-rus, Leptolepis, or ichthyodectiforms, is asmall mobile premaxilla (Patterson, 1977;Patterson and Rosen, 1977).

Although there have been major modi-fications within the Teleostei in the overallshape of the jaw and its component ele-ments, only three major types of changehave occurred in the pattern of intercon-nections in the structural network, of thehead. The first specialization involves ashift in insertion of the mandibulohyoid


ligament to the interoperculum (Fig. 1C:IHL). The interoperculohyoid ligamentcharacterizes the feeding mechanism ofeurypterygian fishes (=Aulopiformes +Myctophiformes + Paracanthopterygii +Acanthopterygii; Rosen, 1973) and effec-tively shifts the action of the hyoid andopercular coupling onto the interopercul-um. Only the interoperculomandibularligament transmits posterodorsal hyoid andopercular movement to the mandible in theEurypterygii, while other teleosts retainthe primitive two-coupling system ofhalecostomes (Fig. IB).

The second major structural specializa-tion within teleosts is the development ofan elongate ascending process on the pre-maxilla and modification of maxillary andpremaxillary articular surfaces and liga-ments, all associated with protrusion of theupper jaw toward the prey during feeding(Fig- 1).

Finally, a number of changes in the jawadductor musculature have occurred.Primitive teleosts are characterized by thepresence of a geniohyoideus muscle ex-tending anteroposteriorly between themandibular symphysis and the ceratohyaland epihyal. The geniohyoideus muscle ofteleosts represents a fused intermandibu-laris posterior and interhyoideus (Fig. 2)of primitive actinopterygians (Winterbot-tom, 1974). Teleosts have lost the bran-chiomandibularis of primitive actinopte-rygians (Lauder, 1980a; Wiley, 1979), aswell as the suborbital adductor component(Fig. 2A: AMa). Only a single unsubdivid-ed lateral adductor muscle is present inprimitive members of the Osteoglossomor-pha, Elopomorpha, and Clupeomorpha,whereas in many euteleostean lineagesboth lateral and medial subdivisions of themain adductor mass (A2/3—Winterbot-tom, 1974) are present. Of particular im-portance for the evolution of protrusilemechanisms in teleosts is the independentevolution in many lineages of one or moreadductor divisions with insertions on themaxilla. Stomiiforms, myctophiforms, someparacanthopterygians, and some primitiveacanthopterygian fishes possess a medialsubdivision of the main adductor mass.Alb, which inserts on the maxilla (Fink

and Weitzman, 1982; Rosen, 1973). Inmany other lineages, a lateral subdivisionof A2/3, Al, inserts tendinously on themaxilla. Based on the diversity of lineageswithin the Acanthopterygii which possessa so-called Al adductor division, it is un-likely that this division is homologousthroughout advanced teleosts, and mus-cular attachments to the maxilla have cer-tainly arisen independently in variousmore primitive teleostean lineages. Anumber of percomorph lineages have asubdivided main adductor mass with sep-arate A2 and A3 components. The A3 usu-ally inserts on the coronomeckelian bonein the Meckelian fossa, while A2 may insertalong the coronoid process and medialface of the dentary and anguloarticular. Awell-developed intramandibular adductordivision (Aw) is present in most teleosts.

Models and mechanisms of upper jawprotrusion: The Acanthopterygii

The ability of many acanthopterygiansto extend the premaxilla and maxilla to-ward the prey during feeding (protrusion)is one of the most widely discussed fea-tures of the teleost feeding mechanism(Alexander, 1967; Eaton, 1935; Gregory,1933; Lauder and Liem, 1981; Liem,1970, 1979, 1980; Nyberg, 1971; Pietsch,1978; Schaeffer and Rosen, 1961; vanDobben, 1937). It is now clear that acan-thopterygians possess a number of differ-ent mechanisms of protrusion involvingnon-homologous articular surfaces, liga-ments, and possibly also muscular controlmechanisms.

Alexander (1967) has provided a me-chanical explanation for premaxillary pro-trusion which seems to apply to someprimitive acanthopterygians. He suggestedthat rotation of the maxilla along its longaxis is caused by depression of the lowerjaw and contraction of the A1 division ofthe adductor mandibulae. Maxillary rota-tion causes the premaxillary process of themaxilla to press against the articular pro-cess of the premaxilla which forces thepremaxilla to protrude anteriorly. The twoprerequisites for this mechanism are (1)maxillary twisting and (2) apposition of thepremaxillarv articular surface with the


premaxillary condyle of the maxilla. Alex-ander (1967) also noted that movements ofthe suspensory apparatus may limit retrac-tion of the protruded premaxilla. Suspen-sory abduction causes the maxillary pro-cess of the palatine to move medially andblock retraction which can only be accom-plished with an adducted suspensorium.Thus suspensorial, maxillary, and premax-illary movements are all coupled with ki-nematic systems involved in mouth open-ing, and mandibular depression ultimatelycontrols protrusion (Liem, 1970).

Liem (1979, 1980) has proposed a "de-coupled model" for cichlid fishes whichlack the requisite anatomical articulationsfor Alexander's (1967) model. This modelinvolves contraction of the epaxial musclesand concomitant stabilization of the pre-maxilla, maxilla, and mandible by contrac-tion of the adductor mandibulae (parts Aland/or A2) and geniohyoideus. By liftingthe neurocranium, jaw protrusion can bemodulated and varied in relation to thedegree of synchronous activity in the ge-niohyoideus and adductor mandibulae. Ex-perimental analysis of prey manipulationin cichlids (Liem, 1979, 1980) shows thatthe epaxial musculature is involved inmodulating upper jaw movement duringmanipulation of prey and that the decou-pled protrusion system allows complex ki-nematic patterns of jaw movement notseen in fishes with coupled mechanisms.

A number of functions have been pro-posed for premaxillary protrusion, but fewof these hypotheses have been tested ex-perimentally. Protrusion may be related tobottom feeding by allowing the mouthopening to be pointed ventrally while thepredator's body remains horizontal (Alex-ander, 1966, 1967). For mid water suctionfeeding, Nyberg (1971), Gosline (1971),and Alexander (1967) have hypothesizedthat protrusion provides an added velocitycomponent to the predator's approach.Nyberg (1971) measured an "added veloc-ity" of 89% of the average attack velocityin Micropterus, and Lauder and Liem(1981) measured an additional velocity of39% of average attack velocity in Lucio-cephalus. It remains to be demonstratedwhether the added velocity resulting from

protrusion actually increases the success ofprey capture, and how jaw protrusion ef-fects the hydrodynamic properties of thefeeding mechanism.

Pharyngeal jaw evolution

Upper pharyngeal jaw dentition inprimitive actinopterygians consists of nu-merous dermal tooth plates aligned with(but not fused to) the pharyngobranchialand epibranchial gill arch elements. Amiaand Lepisosteus are unique among non-te-leost actinopterygians in having manysmall tooth plates grouped into a largeposterior patch so that individual toothplates are not referable to a particular arch(Nelson, 1969). Ventrally, small dermaltooth plates are aligned with a long andslender fifth ceratobranchial (Nelson,1969; Nielsen, 1942). These plates becomeclosely associated in Amia, Lepisosteus, Hio-don, and Elops but are not fused to the fifthceratobranchial.

In the Teleostei, upper pharyngeal den-tition is consolidated into one to five pairedtooth plates (Nelson, 1969). Clupeo-morphs and euteleostean fishes (the Clu-peocephala, Patterson and Rosen, 1977)are derived in having the upper pharyn-geal tooth plates fused to the bony endo-skeletal gill arch elements, and a singlelarge tooth plate fused to ceratobranchialfive. In many euteleostean lineages, onlyone or two large toothplates are present inthe upper pharyngeal dentition (Nelson,1969, p. 492), and pharyngobranchial onecommonly serves as the suspensory ele-ment of the gill basket (see Fig. 5).

A particularly important specializationduring higher teleostean evolution, andone which characterizes the Neoteleostei(Rosen, 1973), is the occurrence of a re-tractor dorsalis muscle. The retractor dor-salis (=retractor arcus branchialium) is apaired muscle which originates on the ver-tebral column (anywhere from the first tothe sixteenth vertebra) and extends an-teroventrally to insert mainly on pharyn-gobranchials three and four (Rosen, 1973;Fig. 5B: RD). A second important inno-vation is the shift in origin of the pharyn-gohyoideus muscle (Fig. 5: PH), whichprimitively originates on ventral gill arch


ESOX

sins'*

FIG. 5. Diagrammatic view of the pharyngeal jawsand their relation to the skull in (A) Esox niger and(B) Ambloplites rupestris. Heavy lines indicate pharyn-geal jaw muscles and their approximate line of action.Thin lines represent the branchial basket. Black barsrepresent electromyographic activity in selected bran-chial muscles during manipulation and deglutition;white bars indicate occasional activity. The short ver-tical line anterior to levator externus one connects thegill basket with the neurocranium and representspharyngobranchial one. Note the two salient special-izations which distinguish advanced euteleosteanssuch as Ambloplites from more primitive forms (Esox):(1) the shift in origin of the pharyngohyoideus muscle(PH) to the urohyal, and (2) the occurrence of a re-tractor dorsalis muscle (RD). Abbreviations: AD5,fifth branchial adductor; GH, geniohyoideus; LE1-4,levator externi muscles; PCi, e, pharyngocleithralisinternus and externus muscles; PH, pharyngohyoi-deus; RD, retractor dorsalis; SH, sternohyoideus.

elements (usually hypobranchial three). Inall myctophiform, paracanthopterygian,and acanthopterygian teleosts (SubsectionCtenosquamata, Rosen, 1973) the pha-ryngohyoideus originates from the uro-hyal (Fig. 5: PH).

The consequence of these two muscular

specializations is the presence in highereuteleosts of mechanical couplings whichincrease the versatility of the pharyngealapparatus. In primitive teleosts (Fig. 5),protraction and retraction of the upperpharyngeal tooth plates must be accom-plished by the levator externus muscles.Levator externus one (and the anterior le-vator interni) have a posteriorly inclinedline of action which results in retraction ofthe upper pharyngeal jaw (Fig. 5A: LEI).Levator externi two and three mediate up-per pharyngeal jaw protraction. In highereuteleosts (Fig. 5B), the retractor dorsalisand antagonistic levator externi three andfour provide a coupling which allows ex-tensive anteroposterior movement of theupper pharyngeal jaws. In addition, thelevator posterior mediates dorsal move-ment of the upper jaws.

Ventrally, the change in origin of thepharyngohyoideus (Fig. 5B) to the urohyalincreases the range of anterior movementof the lower pharyngeal jaw, and, in con-junction with the pharyngocleithralis in-ternus, provides a coupling mediatinganteroposterior excursions.

Experimental evidence from electromy-ographic analyses of pharyngeal muscula-ture in euteleosteans (Fig. 5) reflects theseconstraints on lower pharyngeal jaw move-ment. In Esox, swallowing is accomplishedby synchronous activity in levator externusone and the pharyngocleithralis internus(Fig. 5A: LEI, PCi) indicating that both theupper and lower pharyngeal jaws are re-tracted together. After retraction, levatorexternus three is active, elevating and pro-tracting the upper pharyngeal jaw beforethe second retraction stroke of the lowerjaw.

In Ambloplites (Fig. 5B), the pharyngo-hyoideus and geniohyoideus are activesynchronously to protract the lower pha-ryngeal jaw, while the upper jaw is beingprotracted by the levator externi muscles(Fig. 5B: LE4). Retraction of the upper jaw-occurs via the retractor dorsalis which isactive with the pharyngocleithralis inter-nus (Fig. 5B: RD, PCi). The upper andlower jaws thus are retracted together, al-though the excursion of the lower jaw isonh one-half that of the upper jaw.


In certain acanthopterygian lineages,the lower pharyngeal jaws are fused in themidline and osteological and myologicalspecializations of the entire pharyngeal ap-paratus result in radically different ki-nematic and electromyographic patternsfrom generalized euteleosts (Liem, 1973,1978; Liem and Greenwood, 1981).

CONCLUSIONS

Considerable progress has been madeover the last twenty years in understandingpatterns of structural and functional evo-lution in the feeding mechanism of fishes,and actinopterygians provide an excellentcase study for analyzing the nature andpattern of structural changes in a complexsystem. In particular, an analysis of howthe structural network in the front jaws haschanged (Fig. 1) may be correlated with asimilar analysis for pharyngeal jaw struc-tural and functional systems (Fig. 5) togain an understanding of evolutionary in-teractions between these two coupledstructural features.

One of the major features of the evolu-tion of the actinopterygian feeding mech-anism is an increase in structural complex-ity. As more cladistically derivedmonophyletic lineages of actinopterygiansare considered, structural complexity (de-fined as the number of connections in thestructural network, Fig. 1) increases. Cer-tain changes in the structural network arerelated to structural diversification. For ex-ample, the occurrence of two biomechan-ical pathways mediating jaw opening at thehalecostome level permits structural andfunctional diversification by removingfunctional constraints on the primitivehyoid coupling. A similar relationship be-tween structural diversity and number ofbiomechanical pathways holds for mecha-nisms of jaw protrusion and upper jawmorphology in acanthopterygians (Lau-der, 1981).

In the pharyngeal jaws, three majorstructural and functional modificationshave occurred in teleosts and the distri-bution of these modifications is congruentwith currently accepted monophyletic lin-eages. Consolidation and fusion of pha-ryngeal tooth plates with underlying en-

dochondral gill arch elements characterizesthe Clupeocephala; the presence of a re-tractor dorsalis is shared by neoteleosteanfishes; and the pharyngohyoideus muscleoriginates from the urohyal in the Cteno-squamata. These anatomical specializa-tions result in increased control over preymanipulation and mastication in the phar-ynx and reflect the increasing functionalversatility of pharyngeal jaw elements inactinopterygian evolution.

ACKNOWLEDGMENTS

I gratefully acknowledge the Society ofFellows, Harvard University, for a JuniorFellowship during which this paper wasprepared, and NSF DEB 80-03206 fortravel support. Bill Fink, Sara Fink, andKarel Liem kindly reviewed the manu-script. Figure 2 was drawn by Sara Fink.

REFERENCES

Alexander, R. McN. 1966. The functions and mech-anisms of the protrusible upper jaws of twospecies of cyprinid fish. J. Zool., London149:288-296.

Alexander, R. McN. 1967. The functions and mech-anisms of the protrusible upper jaws of someacanthopterygian fish. J. Zool., London 151:43-64.

Alexander, R. McN. 1970. Mechanics of the feedingaction of various teleost fishes. J. Zool., London162:145-156.

Anker, G. 1974. Morphology and kinetics of thehead of the stickleback, Caiterosleia aculeatus.Trans. Zool. Soc. Lond. 32:311-416.

Ballintijn, C. M., A. van den Burg, and B P. Egber-ink. 1972. An electromyographic study of theadductor mandibulae complex of a free-swim-ming carp (Cyprinus carpio) during feeding. J.Exp. Biol. 57:261-283.

Blot, J. 1978. Origine et phylogenese des poissonsosseux. Boll. Zool. 45:1-21 (Suppl. II).

Dullemeijer, P 1974. Concepts and approaches inanimal morphology Van Gorcum, Assen, TheNetherlands.

Eaton, T. H. 1935. Evolution of the upper jaw mech-anism in teleost fishes. J. Morph. 58:157-172.

Elshoud-Oldenhave, M. J. W. 1979. Prey capture inthe pike-perch, Stizostedion lucioperca (Teleostei,Percidae): A structural and functional analysis.Zoomorphologie 93:1-32.

Elshoud-Oldenhave, M. J. W. and J. W. M. Osse.1976. Functional morphology of the feeding sys-tem in the ruff—Gymnocephalus cernua (L. 1758)—(Teleostei, Percidae). J. Morph. 150:399-422.

Fink, W. L. and S. H. Weitzman. 1982. Relationshipsof the stomiiform fishes (Teleostei), with a de-


scription of Diplophos. Bull. Mus. Comp. Zool.150:31-93.

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

Gosline, W. A. 1971. Functional morphology and clas-sification of teleostean fishes. Univ. Press of Hawaii,Honolulu.

Greenwood, P. H., D. E. Rosen, S. H. Weitzman, andG. S. Myers. 1966. Phyletic studies of teleosteanfishes, with a provisional classification of livingforms. Bull. Am. Mus. Nat. Hist. 131:339-456.

Greenwood, P. H., R. S. Miles, and C. Patterson,(eds.) 1973. Interrelationships of fishes. AcademicPress, London.

Gregory, W. K. 1933. Fish skulls. A study of the evo-lution of natural mechanisms. Trans. Amer. Phil.Soc. 23:75-481.

Grobecker, D. B. and T. W. Pietsch. 1979. High-speed cinematographic evidence for ultrafastfeeding in antennariid anglerfishes. Science205:1161-1162.

Hutchinson, P. 1973. A revision of the redfieldiiformand perleidiform fishes from the Triassic ofBekker's Kraal (South Africa) and Brookvale(New South Wales). Bull. Br. Mus. Nat. Hist.Geol. 22:233-354.

Jessen, H. 1968. Moythomasia nitida Gross und M. cfstrtata Gross, Devonische Paleonisciden aus demOberen Plattenkalk der Bergischen-Gladbach-Paffrather Mulde (Rheinisches schiefergebirge).Palaeontigraphica 128A:87-114.

Lauder, G. V. 1979. Feeding mechanisms in primi-tive teleosts and in the halecomorph fish Amiacalva. J. Zool., London 187:543-578.

Lauder, G. V. 1980a. Evolution of the feeding mech-anism in primitive actinopterygian fishes: Afunctional anatomical analysis of Polypterus, Lep-tsosteus, and Amia. J. Morph. 163:282-317.

Lauder, G. V. 19804. Hydrodynamics of prey cap-ture by teleost fishes. In Biofiuid mechanics, Vol.2, pp. 161-181. Plenum Press, N.Y.

Lauder, G. V. 1980f. The suction feeding mecha-nism in sunfishes (Lepomis): An experimentalanalysis. J. Exp. Biol. 88:49-72.

Lauder, G. V. 1980*. The role of the hyoid appa-ratus in the feeding mechanism of the coelacanthLatimeria chalumnae. Copeia 1980:1-9.

Lauder, G. V. 1980f. On the evolution of the jawadductor musculature in primitive gnathostomefishes. Breviora 460:1-10.

Lauder, G. V. 1981. Form and function: Structuralanalysis in evolutionary morphology. Paleobiol-ogy 7:430^142.

Lauder, G. V. and L. E. Lanyon. 1980. Functionalanatomy of feeding in the bluegill sunfish, Le-pomis macrochirus: In vivo measurement of bonestrain. J. Exp. Biol. 84:33-55.

Lauder, G. V. and K. F. Liem. 1981. Prey capture byLuciocephalus pulcher: Implications for models ofjaw protrusion in teleost fishes. Env. Biol. Fish.6:257-268.

I-auder. G V and S. F. Norton. 1980. As\mmetrKal

muscle activity during feeding in the gar, Lepi-sosleus oculatus. J. Exp. Biol. 84:17-32.

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

Liem, K. F. 1973. Evolutionary strategies and mor-phological innovations: Cichlid pharyngeal jaws.Syst. Zool. 22:425-441.

Liem, K. F. 1978. Modulatory multiplicity in thefunctional repertoire of the feeding mechanismin cichlid fishes. I. Piscivores. J. Morph. 158:323-360.

Liem, K. F. 1979. Modulatory multiplicity in thefeeding mechanism in cichlid fishes, as exempli-fied by the invertebrate pickers of Lake Tangan-yika. J. Zool., London 189:93-125.

Liem, K. F. 1980. Adaptive significance of intra- andinterspecific differences in the feeding reper-toires of cichlid fishes. Amer. Zool. 20:295-314.

Liem, K. F. and P. H. Greenwood. 1981. A functionalapproach to the phylogeny of the pha-ryngognath teleosts. Amer. Zool. 21:83—101.

Liem, K. F. and J. W. M. Osse. 1975. Biological ver-satility, evolution, and food resource exploitationin African cichlid fishes. Amer. Zool. 15:427—454.

Nelson, G. J. 1969. Gill arches and the phylogeny offishes, with notes on the classification of verte-brates. Bull. Am. Mus. Nat. Hist. 141:475-552.

Nielsen, E. 1942. Studies on Triassic fishes from EastGreenland. I. Glaucolepis and Boreosomus. Med.Gron. 138:1-394.

Nyberg, D. W. 1971. Prey capture in the largemouthbass. Amer. Midi. Nat. 86:128-144.

Osse, J. W. M. 1969. Functional morphology of thehead of the perch (Perca fluviatilis L.): An elec-tromyographic study. Neth. J. Zool. 19:289-392.

Osse, J. W. M. and M. Muller. 1981. A model ofsuction feeding in teleostean fishes with some im-plications for ventilation. In M. A. Ali, (ed.), En-vironmental physiology of fishes. Plenum Press, NewYork.

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

Patterson, C. 1977. The contribution of paleontologyto teleostean phylogeny. In M. K. Hecht, P. C.Goody, and B. M. Hecht, (eds.), Major patterns invertebrate evolution, pp. 579-643. Plenum Press,New York.

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. Am. Mus. Nat. Hist. 158:81-172.

Pietsch, T. W. 1978. The feeding mechanism of Sty-lephorus chordatus (Teleostei: Lampridiformes):Functional and ecological implications. Copeia1978:255-262.

Rosen, D. E. 1973. Interrelationships of higher tu-teleostean fishes. In P. H. Greenwood, R. S.


Miles, and C. Patterson (eds.). Interrelationships offishes, pp. 397—513. Academic Press, London.

Rosen, D. E. 1981. Teleostean interrelationships,morphological function and evolutionary infer-ence. Amer. Zool. 22:261-273.

Saint-Seine, P. 1956. L'evolution des actinoptery-giens. Coll. int. Paleont. CNRS Paris 60:27-33.

Schaeffer, B. and D. E. Rosen. 1961. Major adaptivelevels in the evolution of the actinopterygianfeeding mechanism. Amer. Zool. 1:187-204.

Tchernavin, V. V. 1953. The feeding mechanisms of adeep sea fish Chauliodus sloani Schneider. BritishMuseum, N. H., London.

van Dobben, W. H. 1937. Uber den Kiefermechan-

ismus der Knochenfishe. Arch. Neerl. Zool. 2:1-72.

Vandewalle, P. 1979. Etude cinematographique etelectromyographique des mouvements respira-toires chez trois cyprins, Gobto gobto (L.), Barbusbarbus (L.) et Leuciscus leuciscus (L.). Cybium 6:3—28.

Wiley, E. O. 1979. Ventral gill arch muscles and theinterrelationships of gnathostomes, with a newclassification of the vertebrata. Zool. J. Linn. Soc.67:149-179.

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

Documents

Patterns of Evolution in the Feeding Mechanism of Actinopterygian