Evolution of the Feeding Mechanism in Primitive Actionopterygian Fishes: A

JOURNAL OF MORPHOLOGY 163:283-317 (1980)

Evolution of the Feeding Mechanism in Primitive Actinopterygian Fishes: A Functional Anatomical Analysis of Polypterus, Lepisosteus, and Amia

GEORGE V. LAUDER, JR. The Museum of Comparative Zoology, Harvard University, Cambridge, Massachussetts 02138

-4BSTRACT The comparative functional anatomy of feeding in Polypterus senegalus, Lepisosteus oculatus, and Amia calva, three primitive actinopterygian fishes, was studied by high-speed cinematography (200 frames per second) synchronized with electromyographic recordings of cranial muscle activity. Several characters of the feeding mechanism have been identified as primitive for actinopterygian fishes: 1) Mandibular depression is mediated by the sternohyoideus muscle via the hyoid apparatus and mandibulohyoid ligament. 2) The obliquus inferioris and sternohyoideus muscles exhibit synchronous activity at the onset of the expansive phase ofjaw movement. 3) Activity in the adductor operculi occurs in a double burst pattern-an initial burst at the onset of the expansive phase, followed by a burst after the jaws have closed. 4) A median septum divides the sternohyoideus muscle into right and left halves which are asymmetrically active during chewing and manipulation of prey. 5) Peak hyoid depression occurs only after peak gape has been reached and the hyoid apparatus remains depressed after the jaws have closed. 6) The neurocranium is elevated by the epaxial muscles during the expansive phase. 7) The adductor mandibulae complex is divided into three major sections-an anterior (suborbital) division, a medial division, and a posterolateral division. In Polypterus, the initial strike lasts from 60 to 125 msec, and no temporal overlap in muscle activity occurs between muscles active at the onset of the expansive phase (sternohyoideus, obliquus superioris, epaxial muscles) and the jaw adductors of the compressive phase. In Lepisosteus, the strike is extremely rapid, often occurring in as little as 20 msec. All cranial muscles become active within 10 msec of each other, and there is extensive overlap in muscle activity periods. Two biomechanically independent mechanisms mediate mandibular depression in Amia, and this duality in mouth-opening couplings is a shared feature of the halecostome fishes. Mandibular depression by hyoid retraction, and intermandibular musculature, consisting of an intermandibularis posterior and interhyoideus, are hypothesized to be primitive for the Teleostomi.

The Actinopterygii or ray-finned fishes is by that of Schaeffer and Rosen ('61). They exam- far the most diverse group of vertebrates; it ined the major adaptive levels in the evolution includes more than 23,000 fossil and recent of the feeding mechanism of ray-finned fishes species. This tremendous diversity in number and proposed hypotheses ofjaw function at each is mirrored by their extensive morphological level. Their analysis was, of necessity, gradal. and behavioral variations in prey-capture Within the last decade, functional anat- mechanisms and strategies. Diversity in jaw omists have begun to analyze experimentally morphology is a hallmark of the actinoptery- the feeding mechanisms of fishes and to test the gian radiation. hypotheses of earlier investigators who had To date the major study of feeding mecha- suggested possible patterns of jaw movement nisms in primitive actinopterygian fishes is based on post mortem manipulations (e.g.,

0362-252518011633-0283$05.400 1980 ALAN R. LISS, INC. 284 GEORGE V. LAUDER, JR.

Alexander, '66, '67; van Dobben, '35; Gunther The fishes were anesthetized during implanta- and Deckert, '53; Holmquist, 'lo; Kampf, '61; tion with tricaine methane sulfonate, 20f.X-400 Kirchhoff, '58; Tchernavin, '53). Experimental mg/liter. Up to twelve pairs of electrodes were studies have involved high-speed cinematog- implanted at one time, although only five raphy (e.g., Dutta, '68; Liem, '67b; '70; Lauder, channels could be synchronously recorded dur- '79; Nyberg, '7 l), electromyographic analyses ing an experiment. This allowed multiple com- of cranial muscle activity (Ballintijn et al., '72; binations of muscles to be investigated during Elshoud-Oldenhave and Osse, '76; Lauder and any given recording session by simply chang- Liem, '80; Liem, '73, '78; Liem and Osse, '75; ing leads at the connector. A dental drill was Osse, '691, measurement of buccal cavity pres- used to drill small (1 mm diameter) holes in the sures (Alexander, '69; '70; Liem, '78), and strain dermal bones overlying many of the cranial gauge analysis (Lauder and Lanyon, in press). muscles (especially inLepisosteus and Amia) to Almost all of these studies have focused on feed- allow access to the muscle belly. The electrodes ing in advanced teleost fishes, while primitive were glued together into a thin cable, color members of the Teleostei and other actino- coded, and attached to a small plastic clamp pterygian groups have been largely neglected. (visible in Figs. 6, 10). The clamp either was An analysis of the functional anatomy of anchored to a wire passing just dorsal to the feeding in primitive ray-finned fishes is of par- epaxial musculature but ventral to the dorsal ticular importance in providing comparative scales (Lepisosteus, Polypterus), or was screwed data on fish feeding mechanisms and in aiding directly to the skull (Amia). Electrodes were in morphological interpretations of fossil implanted for up to three weeks and, although fishes. Such observations may support infer- no deterioration in signal quality was noticed ence of function from structure in extinct taxa, in this period, the electrodes usually worked and will allow a reinterpretation of the func- loose after two weeks. tional and phylogenetic significance of several During recording sessions, the color-coded features in the earliest ray-finned fishes. Ex- electrode leads were attached to a slip-ring, amination of primitive taxa may reveal charac- rotating connector, that was located just above ters that corroborate (or refute) current the water surface and connected to Gould- hypotheses (Gardiner and Bartram, '77; Brush biomedical amplifiers. Electrical signals Greenwood et al., '73; Liem and Lauder, in were stored on a Honeywell 5600 tape recorder press; Patterson and Rosen, '77; Wiley, '79) of at 37.5 cmlsec and played back for analysis at the systematic relationships of actinoptery- 4.7 cmlsec through a Gould 260 strip chart gians (see Table 1 for example). recorder. This paper focuses on a comparative func- High-speed films (200 frames per second) tional analysis of feeding in Polypterus sene- were synchronized with the electromyographic galus, Lepisosteus oculatus, and Amia calua, recordings by a special synchronization unit three primitive living actinopterygian fishes. that counted pulses from the high-speed cam- "Primitive actinopterygian" as used here refers era and placed a series of coded pulses (see Figs. loosely to all non-teleost actinopterygians. Al- 6, 10, 13: SYN) onto the tape recorder. The though there has been some debate during the films were taken with a Photosonics 16-1PL last century over the phylogenetic position of camera on Kodak 4X Reversal and 4X Negative Polypterus, most of the recent evidence indi- film. Three 600W Smith-Victor filming lights cates that it is an actinopterygian (Gardiner, provided illumination for filming with a shut- '73; Liem and Lauder, in press; Wiley, '79). ter speed of 111200 second. A short period of training (two weeks) was usually necessary to MATERIALS AND METHODS accustom the fishes to feed with the lights on. More than 150 feeding sequences were exam- All fishes studied experimentally were ined for Lepisosteus (4 specimens, including caught in the wild and obtained through com- MCZ 54289,542911,125 forAmia (2 specimens, mercial suppliers. Fish were acclimated to lab- including MCZ 542871, and 100 for Polypterus oratory water and temperature (25°C) for sev- (2 specimens, including MCZ 54290), all taken eral weeks before experimental work was over a two-year period. Polypterus was fed meal begun; each fish was housed in a separate 80 worms (Tenebrio), whereas Amia was fed liter tank. pieces of smelt (Osmerus) and live (occasionally Fine-wire bipolar electrodes (Evenohm S) lightly anesthetized) goldfish (Carassius au- were implanted in the cranial muscles using ratus). Lepisosteus was fed goldfish exclusively. the technique of Basmajian and Stecko ('62). Comparative anatomical observations were ACTINOPTERYGIAN FEEDING MECHANISMS 285 made on the following specimens: Lepidosiren aspect of the neurocranium. In Lepisosteus and paradoxa, U. S. National Museum 162504, one Amia, the premaxillae attach to the neuro- specimen; Latimeria chalumnae, American cranium by elongate "nasal processes" (Patter- Museum of Natural History 32949, one cleared son, '731, whereas in Polypterus, the premaxil- and stained embryo; five uncatalogued Amia lae are fixed to the neurocranium by internal calua, (MCZ); Amia calua, MCZ 8970, three connections to the parasphenoid and by small specimens; Lepisosteus oculatus, MCZ 34650, lateral bony processes extending dorsally on two specimens; three uncatalogued Lepisosteus either side of the ethmoid region (ascending oculatus (MCZ); Polypterus senegalus, MCZ process of Allis, '22). These processes are not 48572, two specimens;Polyodon spathula, MCZ homologous to the ascending processes of the 40481, one specimen; Acipenser huso, MCZ premaxillae in teleosts or to the "nasal proc- 54269, one specimen. esses" in Lepisosteus and Amia. Wiley ('76) Physiological properties of the jaw muscles suggested that these processes in Lepisosteus are of considerable importance in interpreting are not homologous to those in Amia. the electromyographic data. However, only The maxilla in Polypterus forms a rigid part preliminary data are available for actinoptery- of the upper jaw and is attached anteriorly to gian jaw muscles (Lauder and Hylander, un- the premaxilla and lacrimal, and posteriorly to published). In the gar Lepisosteus, the twitch the ectopterygoid. Both the maxilla and ecto- time to peak tension averages 34 msec in the pterygoid bear teeth. The maxilla of Lepis- adductor mandibulae and the latency between osteus is rudimentary (Wiley, '76) and a series the stimulus and the onset of tension is 4 msec. of toothed infraorbital bones form the lateral Tetanic time to peak tension averaged 75 msec toothed margin of the upperjaw. The maxilla of but the time to 50% response was 12 msec. Amia (Fig. 3: MX) shares two features with the primitive teleost maxilla. 1) It attaches to the RESULTS upper jaw by a medial, peglike process which Anatomy fits into a socket formed by the vomer, ethmoid, The cranial anatomy of Polypterus was de- and palatine. This articulation allows the scribed by Allis ('19a, '19b, '221, Luther ('13), maxilla to swing anteroposteriorly in the para- Edgeworth ('35) and Nelson ('69). The cranial sagittal plane (Lauder, '79). 2) Distally, the anatomy of Amia has been monographed by maxilla is connected to the coronoid process of Allis (1897). The anatomy of Lepisosteus has the mandible by a flat maxillomandibular not been well described. Allis ('201, Mayhew ligament. A thin sheet of connective tissue ex- ('241, Patterson ('73), and Wiley ('76) partially tends from the posterior edge of the maxilla to described the cranial osteology, and Luther attach to the infraorbital bones and merges ('13),Edgeworth ('35) and Wiley ('76) provided with the fascia covering the jaw adductor mus- brief accounts of the cranial myology. culature laterally. A ligament from the apex of Those aspects of the cephalic musculo- the coronoid process of the lower jaw (Fig. 3: skeletal system that are necessary for under- MML) extends anteriorly to attach to the me- standing the subsequent functional analysis dial surface of the maxilla. A small supra- are here emphasized, as are features inade- maxilla is present. quately described previously. The musculo- In Amia, Polypterus, and Lepisosteus, the skeletal couplings mediating jaw movements epaxial muscles (Figs. 1, 3: EM) insert on the are discussed and organized within the posterodorsal margin of the neurocranium. The mechanical units (Gans, '69) of the primitive obliquus superioris (Fig. 3: OBS) is a division of actinopterygian skull, as follows: 1) neuro- the hypaxialis and always extends anterodor- cranium and premaxilla, 2) maxilla, sally to insert on the posterolateral margin of 3) suspensory apparatus, 4) mandible, the skull. 5)opercular series, 6) hyoid apparatus, and 7) pectoral girdle. Suspensory apparatus Extensive synonymies for most muscles de- The suspensory apparatus articulates with scribed below may be found in Winterbottom the neurocranium via two articular surfaces- ('74) and Wiley ('79). the craniohyomandibular articulation posteriorly, and the articulation between the Neurocranium and premaxilla; maxilla autopalatine (palatopterygoid in Lepisosteus) In primitive actinopterygians, the pre- and the ethmovomerine region (premaxilla and maxilla forms part of the neurocranial mechan- vomer in Lepisosteus) anteriorly. The orienta- ical unit and is firmly attached to the anterior tion of the axis of these two articulations is 286 GEORGE V. LAUDER, JR. nearly horizontal in Lepisosteus, but oblique in 2: POs, Pop). The palatomandibularis major Amia; in Polypterus the suspensorial axis is extends from a wide origin on the ectopterygoid steeply inclined in a posterodorsal to antero- anterior to the origin of the palatomandibularis ventral direction. The suspensory apparatus minor to insert on the surangular and articu- pivots around the articular axis in a medio- lar. lateral direction. Lateral mobility increases in Parts 3 and 4 of the levator maxillae superi- the serieslepisosteus, Polypterus, Amia as does oris form the anterior adductor division in the moment arm of the quadrate about the ar- Amia (Fig. 3: LMS 3, LMS 4). Levator maxillae ticular axis of the suspensorium with the superioris 4 is cylindrical in shape and neurocranium. parallel-fibered. It originates on the antorbital The suspensory apparatus supports the and extends posteriorly to insert on the pal- adductor musculature of the lower jaw. This atine (Fig. 3: PAL). A few of the more posterior muscle mass is subdivided into anterior, me- fibers reach the ectopterygoid. Levator maxil- dial, and posterolateral parts (homologies are lae superioris 3 is a flat, parallel-fibered muscle given in Table 11, see Discussion). that originates from the palatine and cartilage Anterior divisions of the adductor man- between the palatine and entopterygoid; it ex- dibulae are absent in Polypterus. In Lepis- tends posteroventrally to join the tendon of the osteus, the so-called palatomandibularis minor posterolateral division of the adductor man- (Fig. 2: PMm) and palatomandibularis major dibulae. (Edgeworth, ’35; after Luther, ’13) form the an- The medial division of the adductor man- terior adductor subdivision. The palatoman- dibulae is represented in Polypterus by the so- dibularis minor originates from the dorso- called “pterygoideus” and “temporalis” mus- lateral surface of the ectopterygoid and inserts cles (Allis, ’22) (Figs. 1, 4A AMP, AMt). The on the surangular bone of the lower jaw. This “temporalis”originates from the sphenotic and muscle, actually separable into two well de- frontal and extends posteroventrally where it fined fiber bundles, wraps laterally around the contracts to a wide tendon which inserts medial anterior fibers of the preorbitalis muscles (Fig. to the coronoid process in the mandibular fossa.

ABBREVIATIONS AAM, anterior division of adductor HHI, hyohyoideus inferioris OBS (r, 11, obliquus superioris mandibulae complex HHS, hyohyoideus superioris (right and left sides) AAP,adductor arcus palatini HY, hypaxial muscles (= obliquus OHY, operculohyomandibular AHY,adductor hyomandibulae inferioris) articulation AOP, adductor operculi HYM, hyomandibula OM, operculo-gular membrane AM,AM2,AM2‘, AM2”, IH,interhyoideus OP, operculum posterolateral division of adductor MY, interhyal PA, parietal mandibulae IML, interoperculomandibular PAL, palatine AMP, “pterygoideus” division of ligament PG, pectoral girdle adductor mandibulae complex Ma, intermandibularis anterior PMm, palatomandibularis minor AMt, “temporalis” division of IMP, intermandibularis posterior Pop, preorbitalis profundus adductor mandibulae complex IOP, interoperculum PS, parasphenoid BB, basibranchial LAP, levator arcus palatini Q, quadrate BC, buccal cavity LIM, interoperculomandibular S, splenial BM, branchiomandibularis ligament SH (SHr, SHI), sternohyoideus BML, branchiostego-mandibular LMH, mandibulohyoid ligament (right and left halves) ligament LMS 1-4, levator maxillae SHr (1) ant, SHr (1) post, anterior BSTG, branchiostegal ray(s) superioris division of adductor and posterior fibers of the right CB, ceratobranchial mandibulae complex (left) halves of the sternohyoideus CC, cranial cavity LOP, levator operculi SHT, tendon of sternohyoideus CH, ceratohyal LOPT, tendon of levator operculi SHS, median septum of the CL, cleithrum MAM, medial division of the sternohyoideus CLAV, clavicle adductor mandibulae complex SO, suborbital bone CN, chondrocranium MD, mandible SOP, suboperculum DO, dilator operculi MHL, mandibulohyoid ligament SUS, suspensory apparatus EH, epihyal MML, ligament from the coronoid SYN, synchronization pulse EM, EP, epaxial musculature process of the mandible to the TAM, tendon of adductor GA, gill arch maxilla mandibulae GF, gill filaments MX, maxilla UBSTG, uppermost expanded GP, gular plate NC, neurocranium branchiostegal ray HH, hypohyal OBI, obliquus inferioris ACTINOFTERYGIAN FEEDING MECHANISMS 287 A

BM- 288 GEORGE V. LAUDER, JR. A

PG Fig. 2. Lepisosteus ocnlatus. Lateral view (A) and ventral view (B) after removal of the eye. ACTINOmRYGIAN FEEDING MECHANISMS 289 The “pterygoideus” originates from the para- lateral faces of the ascending process (Fig. 4C: sphenoid and sphenoid (Allis, ’22) and extends LAP). posteroventrally to insert on the medial surface Medial movement of the suspensory ap- of the wide “temporalis” tendon. paratus is accomplished by the adductor arcus In Lepisosteus, the medial adductor division palatini and adductor hyomandibulae (Fig. 5: is represented by the preorbitalis superficialis AAP, AHY). Polypterus lacks an adductor and profundus muscles (Figs. 2,4B: Pop, POs). arcus palatini (defined as fibers extending onto The preorbitalis superficialis originates from a the suspensorium anterior to the hyoman- wide area on the ventral surface of the frontal dibula) and, thus, medial suspensorial move- and dermosphenotic, and runs anteroventrally ment occurs via the adductor hyomandibulae after passing above the orbit to insert muscu- alone. The latter originates on the opisthotic lously on the surangular and articular. The and extends laterally to insert on the medial preorbitalis profundus originates medial to the surface of the hyomandibula. In Lepisosteus, levator arcus palatini on the lateral wall of the the adductor arcus palatini and adductor hyo- braincase and extends anteriorly, medial to the mandibulae form a continuous sheet of eye, before inserting on the articular and sur- parallel-fibered muscle that originates from angular deep to the palatomandibularisminor. the otic cartilage dorsal to the prootic and ex- Parts 1 and 2 of the levator maxillae superi- tends laterally to insert on the medial face of oris (Figs. 3,4C: LMS 1,2)comprise the medial the hyomandibula, metapterygoid, and ento- adductor division in Amia. These muscles arise pterygoid. The adductor arcus palatini is ab- from a common origin on the parasphenoid and sent in Amia. The parallel-fibered adductor anterior face of the hyomandibula and insert hyomandibulae is partially divided by a sep- via separate tendons onto the tendinous inser- tum into anterior and posterior portions, both tions of the medial fibers of two of the pos- of which originate on the prootic and intercalar terolateral adductor divisions, AM2’ ’ and and insert on the medial surface of the hyo- AM2“‘. mandibula. The posterolateral division of the adductor mandibulae in Polypterus has two components, Opercular series a much larger lateral mass (Figs. 1,4A: AM21 The opercular series in Polypterus is com- and a smaller medial one arising from the posed of a large operculum and a small sub- quadrate and inserting into the mandibular opercular bone attached to the anteroventral fossa. The lateral fibers (AM21 arise from the margin of the operculum (Fig. 1: OP, SOP). A connective tissue medial to the suborbital large opercular membrane (Fig. 1:OM) extends bones, the hyomandibula, and connective tis- ventrally and anteriorly into the interman- sue covering the lateral surface of the dilator dibular region. Branchiostegal rays and a operculi. The lateral fibers of this muscle run in levator operculi muscle are absent. a posterodorsal to anteroventral direction, but The adductor operculi (Figs. 1,5:AOP) is con- the medial fibers may extend more antero- tinuous with the adductor hyomandibulae and posteriorly, especially along the ventral mar- has a similar origin. Insertion of the adductor gin of the muscle. Insertion is onto the derm- operculi is posterodorsal to the operculo-hyo- articular (Allis, ’22) and lateral surface of the mandibular articulation along the medial edge splenial. of the dorsal opercular margin, Lateral movement of the suspensorium is The dilator operculi is a long, parallel-fibered mediated by the levator arcus palatini (Figs. muscle (Figs. 1,4A:DO) that originates on the 1-4: LAP). In Polypterus, this muscle arises postfrontal (Allis, ’22) and stretches postero- from the postfrontal (Allis, ’22)with the muscle laterally to attach to the anterodorsal margin fibers fanning out ventrally over the lateral of the operculum. The dilator operculi passes face of the suspensory apparatus. The levator medial to the dorsal fibers of the adductor man- arcus palatini of Lepisosteus has a broad origin dibulae and forms the lateral wall of the from the frontal and dermopterotic, and is spiracular canal. pyramidal in shape with an apical insertion In Lepisosteus, the opercular series consists restricted to a narrow strip of the ectopterygoid of an operculum and a suboperculum. Three and metapterygoid medial to the adductor elongate branchiostegal rays are present. The mandibulae (Fig. 2: AM). Anteriorly, the inser- adductor operculi is continuous with the tion of this muscle becomes musculous. InAmia adductor hyomandibulae and adductor arcus this muscle originates from the sphenotic (= palatini (Fig. 5B: AOP) and has a similar postorbital ossification of Allis, ’22) and inserts origin. Insertion is just posterodorsal to the ar- on the metapterygoid, both on the medial and ticulation of the operculum with the hyoman- 290 GEORGE V. LAUDER, JR.

UBSTG B BSTG

CH ~ IOP

Fig. 3. Amia calm. Lateral view (A) and ventral view (B) after removal of the eye, dorsal aspect of the preoperculum, and gular plate. A cc B , POS AMt \ \ ,PS

-sus -BB -CB

-IH P / \ \ \lU I, I SHS \\SHT SH

Icm

SH' i./ SHS

Fig. 4. Semi-diagrammatic cross-section through the head of (A)Polypterus senegalus, (B)Lepisosteus oculatus, and (C) Amm calua. A 8 AYP

BSTG ‘EH C

Fig. 5. Medial view of the posterior aspect ofthe suspensory apparatus of WPolypterus senegalus,(B)Lepisosteus oculatw, and (C)Amiacalm. Note the relationships of the adductor operculi (AOP)to the adductor hyomandibulae (AHY) and the levator operculi (LOP: in Amia only). ACTINOPTERYGIAN FEEDING MECHANISMS 293 dibula. The dilator operculi (Fig. 2: DO) origi- internal surface and extending posteriorly to nates on the sphenotic and is a short, parallel- insert on the third hypobranchials. In Amia, fibered muscle inserting on the anterodorsal the branchiomandibularis (Fig. 3: BM) is margin of the operculum. A levator operculi is paired at its origin, which is identical to that of absent. Polypterus, but may merge into one median The opercular series of Amia consists of the muscle mass before dividing again to insert on operculum, suboperculum, and interoperculum the third hypobranchials of each side. (Fig. 3: OP, SOP, IOP), as well as the expanded uppermost branchiostegal ray (Fig. 3: UBSTG), Hyoid apparatus which is functionally a part of the opercular The hyoid apparatus is composed of a cerato- apparatus. This branchiostegal ray is attached hyal, an epihyal, and a small styliform inter- to both the ceratohyal and opercular series by hyal which suspends the epihyal from the sus- dense connective tissue. A ligament extends pensory apparatus (Figs. 4,5:CH, EH, IHY). In between the anteroventral aspect of the Polypterus, the interhyal is a small cartilage- uppermost branchiostegal ray and the retro- nous element covered medially by dense con- articular process of the mandible (Fig. 3: BML), nective tissue between the ceratohyal and sus- and a well-defined interoperculomandibular pensorium. ligament connects the interoperculum to the The interhyoideus (Figs. 1-5: IH) arises from retroarticular process (Fig. 3: IML). the ceratohyal and extends anteriorly ventral The adductor operculi is completely sepa- to the gular plate and intermandibularis pos- rated from the adductor hyomandibulae (Fig. 5: terior to meet its antimere in a median raphe. AOP, AHY) and is continuous at its insertion The lateral fibers fan out onto connective tissue on the operculum with the levator operculi dorsal to the intermandibularis posterior. (Fig. 5: LOP). Together, the insertion of these The hyohyoideus inferioris in Polypterus is muscles occupies a broad area on the medial represented by muscle fibers in the expanded surface of the operculum. The origin of the opercular membrane. The hyohyoideus superi- adductor operculi is similar to that of the oris (Fig. 5: HHS) originates on the medial sur- adductor hyomandibulae; both originate from face of the operculum and runs ventrally in the the intercalar, whereas the levator operculi opercular membrane, where the fibers merge originates tendinously from the ventral surface with those of the hyohyoideus inferioris. of the posttemporal and posterior margin of the In Lepisosteus, the hyohyoideus superioris parietal (Fig. 3: LOFT). has a similar origin and also is continuous with The dilator operculi originates on the dermo- the fibers of the hyohyoideus inferioris (Fig. 2: sphenotic and extends posterolaterally. It HHI) which meet in a median raphe ventrally. passes lateral to the hyomandibula, but medial The hyohyoideus inferioris of Amia is a well to the dorsal margin of the preoperculum, to developed muscle that inserts anteriorly on insert on the operculum. both the right and left hypohyals via small tendons (Fig. 3: HHI) and originates on the cerato- Mandible hyal and inner surfaces of the anterior branchiostegal rays. The right and left tendons The mandible in Polypterus, Lepisosteus, and of insertion may be separate, or they may insert Amia is composed of a number of bony elements on the ventral hypohyal of each side in a com- all of which are immovably joined (Figs. 1-4: mon tendon. MD); they form the mandibular mechanical A thick mandibulohyoid ligament extends unit. The intermandibularis posterior (Figs. between the hyoid arch (epihyal; ceratohyal in 1-3: IMP) has a central median raphe and ex- Polypterus) and the retroarticular process of tends between the two mandibular rami. In the jaw (Figs. 1-5: MHL). Amia and Polypterus, the gular plate lies ventral to the intermandibularis posterior, but the Pectoral girdle muscle fibers do not insert onto the plate. A The pectoral girdle of primitive actinoptery- small intermandibularis anterior (Fig. 3: IMa) gians can be treated as a mechanical unit that lies between the mandibular rami ofdmia,an- serves as the origin for the sternohyoideus and terior to the intermandibularis posterior. insertion for the obliquus inferioris. The The branchiomandibularis (= coracoman- sternohyoideus is divided into right and left dibularis, Wiley, '79) in Polypterus (Figs. 1, 4: halves by a median septum that completely BM) is paired throughout its length, origi- divides the muscle (Fig. 4: SH, SHS). InPolyp- nating near the mandibular symphysis on the terus, the sternohyoideus shows three trans- 294 GEORGE V. LAUDER, JR. verse divisions by two transverse septa; the gape (Fig. 7A) occurs both by elevation of the right and left halves of the muscle are com- cranium and by mandibular depression medi- pletely separated for much of their course. The ated by posterodorsal movement of the hyoid sternohyoideus inserts onto the hypohyals and apparatus. This hyoid retraction, initiated by a ventral tendinous thickening (Fig. 4A: SHT) the sternohyoideus, is transmitted to the extends anteriorly onto the ventral surface of posteroventral aspect of the mandible by the the fleshy tongue. In Amia, the sternohyoideus mandibulohyoid ligament (Fig. 1: MHL) and is also divided transversely by two septa and results in mandibular depression. Hyoid de- inserts onto the hypohyals. pression reaches its maximum from 10 to 25 No transverse septa were found in the msec after peak gape (Fig. 7A). There is a con- sternohyoideus of Lepisosteus and the sterno- sistent 15-25 msec delay between the onset of hyoideus inserts onto the hypohyals by two activity in the sternohyoideus and the initia- stout tendons (Fig. 2: SHT). A thickened, ten- tion of mandibular depression (Fig. 6). dinous expansion (Fig. 4: SHT) extends an- The adductor operculi becomes active 10 teriorly onto the ventral surface of the basihyal msec after the onset of activity in the sternohy- tooth plates. oideus, and the adductor mandibulae (parttwo) The obliquus inferioris, a division of the frequently also displays low-level activity at hypaxialis (Winterbottom, ’741, is composed of this time (Fig. 6: AOP, AM2).Opercular adduc- fibers extending anteroventrally. These fibers tion (Fig. 7A) is characteristic of the expansive insert on the pectoral girdle. There is no con- phase, peak adduction usually occurring either tinuity between the obliquus inferioris and the synchronously with, or slightly before, peak sternohyoideus fibers. gape. Late in the expansive phase the levator arcus palatini fires (Fig. 6: LAP), although sus- FUNCTIONAL ANALYSIS OF PREY CAPURE pensory abduction does not reach a maximum Jaw movements during prey capture in until 10 to 20 msec after peak gape. fishes can be divided into three phases-a pre- Activity in the hyohyoideus superioris is paratory phase, an expansive phase, and a variable, but during vigorous strikes this mus- compressive phase. To date preparatory phases cle is active from 10 to 15msec before activity in have only been observed in acanthopterygian the sternohyoideus and remains active for up to teleosts. No preparatory phase was recorded in 30 msec. Polypterus, Lepisosteus, or Amia. The expansive phase is defined here as the time from the Compressive phase start of mouth opening to peak gape. The com- The compressive phase is initiated by high- pressive phase extends from peak gape to com- level activity in the dilator operculi (Fig. 6: DO, plete closure of the jaws. frame 141, which initiates opercular abduction (Fig. 7A: frames 1Z22).Fifteen to twenty msec Polypterus senegalus after the onset of the compressive phase, the Expansive phase adductor mandibulae part two becomes active The duration of this phase varies considera- (Fig. 6: AM21 and within 10 msec, a sharp burst bly from rapid (60 msec) to slow (125 msec) of activity occurs in the other two divisions of strikes and results in the creation of a flow of the adductor complex, the “temporalis” and water into the mouth cavity, enabling prey to “pterygoideus” (Figs. 1,6:AMP, AMt). In slow be sucked up off the bottom. The complete strikes, no activity in any adductor division is strike may take from 120 to 300 msec. The observed for nearly 100 msec after the onset of onset of the expansive phase is characterized by mouth opening and initial mouth closure ap- synchronous activity in the paired sternohy- pears to occur primarily by elastic recoil in the oideus muscles and obliquus inferioris (Fig. 6: adductor mandibulae complex. SHr, 1; HY). Activity in the epaxial muscles After the end of the compressive phase, a (Fig. 6: EP) begins 5 msec before the onset of second burst of activity occurs in the adductor activity in the sternohyoideus. The increase in operculi and restores the operculum to its ini-

Fig. 6. Polypterus senegalus. Frames 1, 5, 14, and 22 (on left) from a high-speed (200 frames per second) film synchronized with electromyographic recordings of cranial muscle activity (on right). Shutter synchronization pulse (SYN) provided accurate correlation. Peak gape occurs in frame 14. Note the lack of overlap in activity between mouth opening muscles (SHr, 1; HY; EP) and closing muscles (AM2,AMt, AMp). ACTINOPTERYGIAN FEEDING MECHANISMS 295

POLYPTE RUS SE NEGALUS FRAME1 5 14 22 II I I 296 GEORGE V. LAUDER, JR. A

GAPE

DEPRESSION

. . ' ' ' ' ' ' ' ' ' ' '

ABDUCT ION :: :: PERCULARt abduct

6 OPERCULAR ABDUCTION r

'- GAPE

0 """"

HYOID DEPRESSION 1-

Fig. 7. Polypterus senegalus. Graphic representation of gape, hyoid depression, and opercular abduction versus time during (A) the strike and (B) prey ejection during mastication (see text). Ordinate scale is relative. Note especially the changingpattern of opercular movement: adduct-abduct-adduct in the strike and abduct-adduct-abductduring prey ejection. ACTINOFTERYGIAN FEEDING MECHANISMS 297 tial rest position (Figs. 6: AOP; 7A). Occasional movements, the prey is rapidly sucked back in low level activity is seen in the branchioman- to the buccal cavity (Fig. 8B) and the cycle dibularis (Fig. 6: BM), whereas the interman- started again. dibularis posterior and interhyoideus are in- During the ejection phase of mastication, the variably silent. operculum abducts while the mouth is closed Two main features characterize the strike in (Fig. 7B), then adducts as the mouth opens, PoZypterus- 1) the double burst pattern of forcing water out the mouth, and finally ab- activity in the adductor operculi, and 2) the ducts again as the mouth closes. This is the complete lack of overlap between the early ex- reverse of the adduct-abduct-adduct pattern pansive phase muscles (the sternohyoideus, seen during the strike. Depression of the hyoid epaxial, and hypaxial muscles) and activity in clearly is synchronized with gape (Fig. 7B). the adductor complex. Interspersed with prey ejection sequences are strong “crushing” and compressive Mastication and deglutition movements. These involve high-level activity The pattern of muscle activity and jawbone in the branchiomandibularis, intermandibu- movement during mastication of the prey fol- lark posterior, interhyoideus, and adductor lowing capture may vary significantly from operculi (Fig. 8C), all compressive muscles. that seen during the strike. The usual pattern Crushing of prey also may occur with associ- of prey manipulation involves a series of short ated activity in the adductor mandibulae part “spitting out” movements, during each of which two, levator arcus palatini, and dilator oper- the prey is partially ejected from the buccal culi, but no activity occurs in the sternohy- cavity and then crushed between the jaws. oideus (Fig. 8A). Differential activity occurs After a series of these ejection and crushing between the adductor part two, and the “tem-

A B C

AOP

Fig. 8. Polypterus senegalus. Block diagram of three common patterns of cranial muscle activity occurring during mastication. Pattern B most closely corresponds to the pattern seen at the strike. See text for further explanation. 298 GEORGE V. LAUDER,JR. poralis” and “pterygoideus” divisions during sternohyoideus, hyohyoideus superioris, di- mastication. These two divisions often show lator operculi, levator arcus palatini, and ep- nearly synchronous isolated bursts of high- axial muscles, and all show nearly the same level activity (see Fig. 8B: AMP, AMt) inde- relative timing of activity as seen at the strike. pendently from activity in AM2 (Fig. 8A). Throughout the expansive and compressive During positioning of the prey in masticatory phases, low-level activity occasionally is sequences, a distinct pattern of asymmetrical recorded in the branchiomandibularis. activity is often observed between the right and Lepisosteus oculatus left halves of the sternohyoideus. Complete asymmetry occurs (Fig. 9B) in which both the Expansive phase anterior and posterior portions of one side are The initial strike atthe prey byLepisosteus is active simultaneously, while no activity is seen extremely rapid; the complete process of prey in the other half. Anteroposterior asymmetry capture (from the start of mouth opening to also occurs in which the anterior portions of closure of the jaws) often occurs in as little as 20 each side are active independently from the msec. There is remarkably little variation in posterior fibers (Fig. 9C). Finally, asymmetries the duration of the strike and all strikes occur in timing are observed in which one side initi- within a time span of 20-35 msec. Thus, the ates activity 10-50 msec prior to activity in the expansive phase may last only one frame at opposite side (Fig. 9A). This asymmetrical ac- filming speeds of up to 200 frames per second tivity could not be correlated with any obvious (e.g., Fig. 10). features of the pattern of jaw movement, al- Peak gape is rapidly achieved both by de- though movements of the elements most likely pression of the lower jaw and elevation of the to be affected by asymmetrical sternohyoideus cranium and upper jaw (Fig. 11).Mandibular activity-the hypohyals, ceratohyals, and depression is initiated by symmetrical activity basibranchials-could not be directly observed in the paired sternohyoideus muscles (Fig. 10: by light cinematography. SHr, 1) which are active simultaneously with When the prey has reached the front of the the obliquus inferioris (Fig. 10: HY). A delay of jaws as a result of repeated cycles of ejection 5-10 msec occurs between the onset of activity and crushing, it is rapidly sucked back into the in the sternohyoideus muscle and the start of buccal cavity by a sequence of events similar to mandibular depression. The epaxial muscles those of the initial strike at the prey (Fig. 8B). become active 5 msec after the onset of activity The duration of the expansive and compressive in the sternohyoideus (Fig. 10: EP). phases is greater than at the strike, and ac- Prey capture is effected by a lateral move- tivity is generally seen in the branchioman- ment of the head towards the prey as the jaws dibularis, intermandibularis posterior, and open (see Fig. 10). This lateral movement is interhyoideus. The adductor operculi may be- achieved by differential timing of activity become active up to 70 msec before activity in the tween the right and left sides of the obliquus

A 8 C SHr ant 1 II L SHr post I ,--+a SHI ant ,, II I

1 u 100 msec

Fig. 9. Polypterus senegalus. Block diagram of three patterns of asymmetrical activity in the sternohyoideus occurring during mastication. Note bath anteroposterior asymmetry and right-left asymmetry. ACTINOFTERYGIAN FEEDING MECHANISMS 299

LEPISOSTEUS OCULATUS

20 eV OBSL -

Fig. 10. Lepisosteus oculntus. Frames 1, 2, 3, and 4 (on left) from a high-speed (200 frames per second) film synchronized with electromyographic recordings of cranial muscle activity (on right). Peak gape occurs between frames 2 and 3. Note the extemive overlap in activity periods of the cranial muscles and the close synchrony in time of onset of activity. 300 GEORGE V. LAUDER, JR. superioris (Fig. 10: OBSr, 1). This muscle is a does not begin until 20 msec into the expansive division of the hypaxialis (Winterbottom, '74), phase (Fig. 11).The hyohyoideus superioris is and the fibers run anterodorsally to insert on also active at the onset of the expansive phase. the posterolateral area of the otic region (pter- Because components of the adductor man- otic); thus, they effect lateral bending of the dibulae complex become active during the ex- head. In Lepisosteus, the obliquus superioris on pansive phase, it is impossible to distinguish the side of the head towards the prey is active opening from closing muscles on the basis of the 1&15 msec before the opposite side during time of electrical activity. The preorbitalis pro- strikes involving significant lateral head fundus is the first component of the adductor movement (Fig. 10). In strikes in which the complex to become active; the adductor man- prey is caught with little lateral head move- dibulae and preorbitalis superficialis fire 5-10 ment, the differences in timing between sides msec later (Fig. 10: AM, Pop, POs). The become small (< 2 msec). adductor mandibulae is consistently active 10 The adductor operculi becomes active from msec after the onset of activity in the sternohy- up to 5 msec before activity in the sternohyoideus, and in spite of the rapidity of the strike, oideus to 5 msec after (Fig. 10: AOP) and results these two muscles are rarely active simultane- in opercular adduction (Fig. 11) during mouth ously. The interhyoideus (Fig. 10: IH) shows a opening. Although the levator arcus palatini is short high-amplitude burst of activity within 5 active within 5 msec of the sternohyoideus (Fig. msec of activity in the sternohyoideus. 10: LAP), suspensorial and opercular abduction Compressive phase The compressive phase consists of depression NEUROCRANIAL of the upper jaw and mandibular adduction to ELE VAT ION trap the prey between the jaws (Fig. 10: frames 3 and 4; Fig. 11).All muscles active at this time were also active in the expansive phase. 5O~ - Hyoid depression reaches a peak (Fig. 11) well after the jaws have begun to close and at 0- v""" peak gape only a small depression of the hyoid has occurred (Fig. 10: frame 2). Opercular ab- MANDIBULAR duction rises to a sharp peak during the compressive phase (Fig. 11) and the operculum 30120 PRESSI ION slowly returns to its resting position due to a second burst of activity in the adductor operculi (Fig. 10: AOP). A double-burst pattern of activity also is seen in the interhyoideus, preorbitalis profundus, and also occasionally in o'----, " " " '' the sternohyoideus resulting in the mainte- HYOID DEPRESSION nance of a depressed hyoid after the mouth has closed. In general, all muscles become active in the expansive phase and continue with at least :: :: 01 low-level activity throughout the compressive phase. OPERCULAR ABDUCTION Mastication and deglutition Following initial capture of the prey, a complex pattern of manipulatory movements oc- 0 2 4 6 8 10 12 14 16 18 20 curs. This orients the prey so that it may be 'Bswallowed head first. This process results in a number of asymmetrical activity patterns in TIME (in ,005sec) the cephalic and anterior body muscles and will be treated in detail elsewhere (Lauder and Nor- Fig. 11. Lepisosteus oculatus. Graphic representation of ton, in press). neurocranial elevation (in degrees), mandibular depression (in degrees), hyoid depression, and opercular abduction ver- In general, patterns of muscle activity during sus time during a strike. Ordinate scale is relative for hyoid manipulation may be divided into two types. depression and opercular adduction. One pattern is very similar to that of the initial ACTINOFTERYGIAN FEEDING MECHANISMS 301 A B C SH r I I I I SH I n I HHS I AM 1-

POS

POP

LAP u HY I 100 msec AOP

L -~

Fig. 12. Lepisosteus oculatus. Block diagram of three common patterns of muscle activity occurring during mastication. Pattern B most closely corresponds to the muscle activity pattern seen at the strike. Open bars represent occasional activity.

strike (Fig. 12B).Notable differences, however, adductor operculi also may occur during these include the generally continuous, uniphasic ac- manipulations. tivity of the adductor operculi, a delay of up to Manipulations of the second type typically 30 msec in initiation of activity in the levator last 5&100 msec. arcus palatini, and the frequent lack of activity Amia catva in the preorbitalis superficialis. “Crushing” movements also occur during Expansive phase prey manipulation and correlate with high The complete strike inAmia usually lasts for amplitude synchronous activity in the adductor 6CL100 msec and is relatively invarient in mandibulae, preorbitalis superficialis and pro- length, although it is more variable than in fundus, and the levator arcus palatini (Fig. 12: Lepisosteus. The expansive phase is usually AM, POs, Pop). Occasional activity also is seen shorter than the compressive phase, lasting in the interhyoideus at this time. about 20-40 msec, and mandibular depression The second type of muscle activity pattern is initiated by synchronous high amplitude ac- during manipulation involves mainly the ep- tivity in the paired sternohyoideusmuscles and axial muscles and sternohyoideus. Activity in activity in the levator operculi (Fig. 13: SHr, 1, both of these muscles (Fig. 12A) or activity in LOP). The sternohyoideus muscle causes hyoid one side of the sternohyoideusonly (Fig. 12C) is retraction, a movement that is transmitted to correlated with lateral bending of the head and the posteroventral aspect of the mandible by slight mandibular depression (Lauder and Nor- the mandibulohyoid ligament (Fig. 3: MHL). ton, in press). All patterns of sternohyoideus The obliquus inferioris is also active to stabilize asymmetry found in Polypterus (Fig. 9) also the pectoral girdle at this time. Contraction of occur in Lepisosteus. Activity in both sides of the levator operculi results in a dorsal rotation the sternohyoideus, interhyoideus, and of the opercular series which is transmitted to 302 GEORGE V. LAUDER, JR.

AMlA CALVA FRAME 1 4 711

,884 I I, I SHI

II ::: ACTINOPTERYGIAN FEEDING MECHANISMS 303 the mandible by the interoperculomandibular maximum value before peak gape (Fig. 15) and ligament (Fig. 3: IML). A delay of 15 msec oc- depression of the posterior margin of the gular curs between activity in the levator operculi plate results from ventral hyoid movement. and the start of mouth opening. During the expansive phase, the branchio- Opercular adduction takes place during the stegal rays are tightly adducted by the hyo- expansive phase, and abduction of the oper- hyoideus superioris (see Fig. 14A, B). culum begins just before peak gape is reached (Fig. 14A, B; Fig. 15). Adduction results from Compressive phase the first burst of activity in the adductor oper- Although activity in the adductor man- culi (Fig. 13: AOP) and occurs despite consider- dibulae complex starts in the expansive phase, able overlap with activity in the dilator oper- it continues into the compressive phase and culi and levator arcus palatini (Fig. 13: DO, results in adduction of the jaws (Fig. 13: AM2’ ‘, LAP). The suspensory apparatus is only AM2’, LMS4; Fig. 15). The levator arcus pal- slightly abducted during the compressive atini and dilator operculi remain active and phase (Fig. 15). result in a large lateral expansion of the oro- Both muscles dorsal to the gular plate-the branchial cavity (Fig. 14C, D; Fig. 15) as the intermandibularis posterior and inter- mouth closes. hyoideus-show a sharp, high-level initial The second burst of activity in the adductor burst of activity within 5 msec of the onset of operculi occurs in the compressive phase and activity in the sternohyoideus (Fig. 13: IMP, continues subsequent to complete jaw closure IH). The intermandibularis may contribute to (Fig. 13: AOP). The branchiomandibularis also the maintenance of suspensorial adduction exhibits a strong burst of activity after the jaws during the expansive phase, have closed and may function to rapidly com- The branchiomandibularis also shows a press the buccal cavity by protracting the sharp initial burst of activity within 10 msec of branchial basket. sternohyoideus activity (Fig. 13: BM). During the compressive phase, the maxilla All jaw muscles become active during the returns to its rest position, while the gular expansive phase as in Lepisosteus. The levator plate and hyoid remain depressed (Fig. 15). maxillae superioris (parts three and four) are Branchiostegal expansion is maximal at this the first adductor divisions to become active time. (Fig. 13: LMS 4); they initiate activity simultaneous with mouth opening muscles, the Mastication and deglutition sternohyoideus and levator operculi. Ten milli- Vigorous chewing movements follow prey seconds later, the adductor mandibulae (part capture in Amia, unless the prey has been 2”) is activated followed in 10 msec by AM2’ sucked directly into the stomach during the (Fig. 13: AM2‘ ’, AM2’). These differences are strike. The manipulatory movements seem to relatively consistent between different feeding be associated with positioning of the prey just sequences, although occasionally AM2‘ ’ and anterior to the esophagus to facilitate degluti- AM2‘ are active simultaneously. tion. The most obvious feature of the movement Strong compressive movements of the buccal pattern during the expansive phase is the ex- cavity walls occur frequently during “chew- tensive arc through which the maxilla swings ing,” and activity is most commonly seen in the (Fig. 13: frame 4). Maximally, it forms nearly a adductor operculi and branchiomandibularis 90” angle to the body axis and effectively (Fig. 16A: AOP, BM). These are the two mus- occludes the corners of the mouth opening by cles most frequently active during prey manip- creating a nearly round tubular orifice. Al- ulation; often they display repeated simultane- though hyoid depression does not peak in the ous bursts for periods of up to several seconds. expansive phase, it does achieve nearly its Occasionally the interhyoideus may be active

Fig. 13. Amia calua. Frames 1, 4, 7, and 11 (on left) from a high-speed (200 frames per second) film synchronized with electromyographic recordings of cranial muscle activity (on right). Peak gape occurs at frame 4. Note the extensive overlap in muscle activity periods and the relative movement of the predator and the prey during the strike. 304 GEORGE V. LAUDER, JR. ACTINOPTERYGIAN FEEDING MECHANISMS 305

SUSPENSORIAL ABDUCT ION

2t /1 l1.-._-/ n-l1.-._-/ 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 TIME (in ,005 sec)

Fig. 15. Amia calua. Graphic representationof gape, hyoid depression, suspensorial abduction, and opercular abduction during the strike. Ordinate scale is relative.

during these bursts, and the levator maxillae could not be clearly correlated with any specific superioris four (with the sternohyoideus and movement pattern. obliquus inferioris) may show alternating Other manipulatory activity patterns (Fig. bursts of activity (Fig. 16A). This pattern also 16C) demonstrate differential activity between demonstrates differential activity between the AM2" and AM2' components of the parts three and four of the levator maxillae adductor mandibulae complex as well as a dou- superioris. ble burst pattern of activity in the levator arcus The patterns of right-left and anteroposterior palatini (Fig. 16C). No activity in the obliquus asymmetry observed in the sternohyoideus of inferioris or epaxial muscles is observed. Polypterus and Lepisosteus also are found in Pattern B (Fig. 16) represents a manipu- Amia (Fig. 17). Sternohyoideus asymmetry latory sequence with muscle activities similar 306 GEORGE V. LAUDER, JR. A B C

HHS I AMY

AM;

LMS4

LMS3

BMI-- ~____I IMP I

IH I

LAP AOP- I DO - I

~ HY 100 msec

HHI

Fig. 16. Amia calua. Block diagram of three common patterns of muscle activity occurring during mastication. Pattern B most closely corresponds to the muscle activity pattern occurring at the strike. Open bars represent occasional activity.

STRIKE MANIPULATION

Fig. 17. Amia calua. Representative pattern of asymmetrical muscle activity in the sternohyoideus occurring during manipulation and mastication of prey. The symmetrical pattern of activity at the strike during the same experiment is shown to the left. ACTINOF’IXRYGIAN FEEDING MECHANISMS 307 in timing to those seen at the initial strike. No HY, SH), whereas inAmia (Fig. 161, activity of activity is recorded in the branchiomandibu- the obliquus inferioris may slightly precede or laris. The levator maxillae superioris and occur simultaneously with that of the sterno- adductor mandibulae divisions of the jaw hyoideus. Relative stability of the pectoral gir- adductor complex fire simultaneously. The rel- dle is important in providing a fixed origin for ative timing of muscle activity is more variable the sternohyoideus, allowing it to effect motion during prey manipulation than during the of the lower jaw, rather than protraction of the strike. pectoral girdle. DISCUSSION Amia has an additional mechanism mediating mandibular depression. Contraction of the Prey capture: comparisons levator operculi causes a posterodorsal rotation and generalizations of the operculum about the operculohyoman- Polypterus, Lepisosteus, andAmia all depress dibular articulation. This rotation is trans- the mandible by contraction of the sterno- mitted to the mandible via the suboperculum, hyoideus muscle. Sternohyoideus activity re- interoperculum, and the interoperculoman- sults in a posteroventral rotation of the hyoid dibular ligament and causes mandibular de- apparatus. This rotation pivots the interhyal pression. The electromyographic data (Fig. 13: bone (Fig. 5: IHY) which suspends the hyoid LOP, AOP, AM2’ ‘) confirm the hypothesis apparatus from the suspensorium. Rotation of (Lauder, ’79) that Amia should have a levator the hyoid apparatus causes a posterodorsal operculi-opercular series-mandible coupling. force to be applied to the retroarticular process This coupling in Amia is homologous to that of of the mandible via the mandibulohyoid liga- teleosts, and represents a previously unrecog- ment and causes mandibular depression be- nized shared specialization of the halecostome cause the force is applied ventral to the man- fishes (Halecomorphi plus Teleostei; see Table dibular axis of rotation (see Fig. 18). 1). The ligamentous connection between the Generally it is believed that the geniohy- hyoid and mandible permits mandibular de- oideus muscle of teleosts (not homologous to the pression by hyoid retraction; the interhyal bone geniohyoideus of tetrapods) is formed by the greatly increases the mechanical advantage of fusion of the intermandibularis posterior and the hyoid retraction mechanism. If the hyoid interhyoideus muscles of primitive actino- articulated directly with the suspensory ap- pterygians (Winterbottom, ’74). Primitively, paratus, contraction of the sternohyoideus then, two myosepta should occur in the teleost would cause only a limited posterodorsal geniohyoideus: a longitudinal septum dividing movement of the mandibulohyoid ligament be- the right and left sides of this muscle, and a cause the origin of this ligament on the hyoid transverse septum separating the “interman- would have to move around the articulation of dibularis component” from the “interhyoideus the hyoid with the suspensorium. Hyoid de- component.” The innervation of the anterior pression then could produce only a small pos- and posterior portions also differs; the inter- terior displacement of the ligamental origin. mandibularis component is trigeminal, The intercalation of an interhyal bone has whereas the interhyoideus is innervated by the two effects. It increases the effective moment facial nerve. arm of the origin of the mandibulohyoid liga- In Amia, both the intermandibularis pos- ment relative to the suspensorium, and it also terior and interhyoideus are active at the onset introduces a second center of rotation (Fig. 18) of the expansive phase (Fig. 13: IMP, IH). In that allows the hyoid to be retracted relative to Lepisosteus, the interhyoideus is also active in the suspensorium. This greatly increases the the expansive phase (no data are available on excursion induced by the force resulting from the intermandibularis posterior), while in Po- sternohyoideus contraction-i.e., a given in- lypterus, neither muscle is active during the cremental hyoid retraction results in greater strike (Fig. 6). The pattern of activity in Polyp- mandibular depression. terus most closely resembles the timing of During contraction of the sternohyoideus, geniohyoideus activity in primitive teleosts the pectoral girdle is stabilized by activity of (Lauder and Liem, ’80).No activity occurs dur- the obliquus inferioris. In Polypterus and ing the expansive or compressive phases and Lepisosteus, the obliquus inferioris shows a only following closure of the mouth does high amplitude burst of activity simultaneous contraction of the geniohyoideus protract the with activity in the sternohyoideus (Figs. 6,lO: hyoid arch. Both the intermandibularis pos- 308 GEORGE V. LAUDER, JR. terior and interhyoideus have important roles manipulating prey prior to swallowing. Con- in compressing the buccal cavity during chew- traction of the right side of the sternohyoideus, ing (Fig. 8). for example, will exert a posterolateral force on In Amia, activity of the intermandibularis the right hypohyal and the component of this posterior during the expansive phase may con- force, acting perpendicular to the sagittal tribute to suspensorial adduction (Fig. 13: IMP; plane, will cause a lateral movement of the Fig. 15), but the significance of concomitant tongue and anterior portion of the hyoid ap- activity in the interhyoideus and branchio- paratus. Asymmetrical activity was not ob- mandibularis is unclear. These muscles may served in other cephalic muscles, although the directly contribute to mandibular depression obliquus superioris and epaxial muscles by forming a "contractile link between the showed marked asymmetry during certain hyoid, branchial basket, and mandible, or they manipulatory patterns in Lepisosteus (Lauder may play a role in elevating the floor of the and Norton, in press). mouth during the early part of the strike. The branchiomandibularis is the main compressive Suction feeding muscle during chewing. The pattern of activity In both Polypterus and Amia, prey capture in the intermandibularis posterior and the occurs by an expansion of the buccal cavity that interhyoideus of Amia does not correspond to causes a pressure reduction and draws water in known activity patterns in the geniohyoideus through the mouth. The unidirectional flow of of primitive teleosts. fluid is the result of a precise control of the A differentiation in the timing of activity timing of jaw movements. The double-burst between the different components of the pattern of activity in the adductor operculi, for adductor mandibulae complex also has been example, allows water to enter the buccal cav- observed. The medial adductor division may be ity only through the anterior mouth opening. active independent of the posterolateral divi- The hyohyoideus superioris is also active at the sion (Polypterus), or the anterior division may start of the expansive phase and prevents water be active independent of the medial or postero- inflow from the posteroventral margin of the lateral division (Amia). In Amia, the two com- opercular cavity. This timing of activity by the ponents of the anterior division-the levator adductor operculi and hyohyoideus superioris maxillae superioris parts three and four-also is maintained even in Lepisosteus. Hyoid de- may be active independently. Only occasional pression always reaches its peak coincident independent adductor activity was found in with, or after, peak gape and the hyoid remains Lepisosteus; activities of the adductor man- depressed well after the jaws have closed. A dibulae, preorbitalis superficialis and profun- similar pattern is seen in suspensory abduction dus, and the levator arcus palatini usually which peaks after the expansive phase has been overlap nearly completely during chewing se- completed. The delay in hyoid depression and quences (Fig. 12). suspensorial abduction may contribute to an The function of the levator maxillae superi- unidirectional fluid flow through the oro- oris part four in Amia is obscure. The muscle branchial chamber. Water exits through the extends from the antorbital to the palatine and opercular cavity as the mouth closes and oper- ectopterygoid, and has relatively little moment cular dilation is maintained throughout the arm about the axis of suspensorial movement. compressive phase. It may contribute slightly to suspensorial An important characteristic of the suction- adduction. feeding mechanism is the creation of a circular In all three genera, asymmetrical activity orifice through which water is drawn (Lauder, was noted in the sternohyoideus muscle, and '79; Osse, '69). InAmia, this is accomplished by the anterior and posterior fibers also act at dif- the extreme anterior swing of the maxilla, ferent times. Lauder and Norton (in press) cor- which, along with the connective tissue extend- related the activity in one side of the sterno- ing posteriorly to the suborbital bones, serves hyoideus with lateral bending of the head and to occlude the margin of the gape (Fig. 13: slight mandibular depression in Lepisosteus. frame 4). Osse ('76) has measured buccal and This asymmetry plays an important role by opercular cavity pressures in Amia and found manipulating the prey into the preferred orien- average values of -120 and -55 cm H,O, re- tation for swallowing. InAmia and Polypterus, spectively. In Polypterus, the corner of the asymmetrical activity of the sternohyoideus mouth is occluded by the thickened connective may produce lateral movements of the tongue tissues of the lip and cheek (Fig. 6: frame 14). in the buccal cavity and aid in positioning and In a theoretical model of the suction feeding ACTINOPTERYGIAN FEEDING MECHANISMS 309 mechanism, Elshoud-Oldenhave and Osse (‘76) the capture of elusive prey items. Visual input have proposed the following sequence for suc- seems only to determine the time of onset of the tion feeding in fishes: An initial preparatory programmed pattern. phase in which the volume of both the buccal Two hypotheses are consistent with the pat- and opercular cavities is decreased; aPhase Z in tern of muscle activation observed during a which the buccal and opercular cavities are ex- preprogrammed strike, and either may account panded with the mouth closed a Phase ZI when for the resulting movement pattern: 1) A pre- both chambers are rapidly enlarged; and a cise neural program could activate the cranial Phase ZZZ in which the mouth is closed and a muscles in the proper temporal sequence and positive pressure is generated in the buccal cav- with the appropriate preprogrammed patterns ity forcing the water over the gills. of force development. Under this hypothesis the Although this model seems to apply to ad- mouth-opening muscles are activated nearly vanced teleosts, no preparatory phase has been simultaneously with the closing muscles in observed in non-acanthopterygian teleosts and Lepisosteus, but more motor units of the open- none was observed inAmia. In addition, lateral ing muscles are activated at a higher frequency expansion of the suspensory apparatus does not of stimulation for the first 5-10 msec to cause begin until Phase I1 and reaches its peak in mouth opening. Increasing stimulation rate Phase 111. Fishes for which this model does and recruitment of motor units in the closing apply appear to be able to generate signifi- muscles then would result in jaw adduction. cantly greater negative buccal cavity pressures 2) A second hypothesis involves the near (-400 to - 600 cmH,O, Alexander, ’70; Liem, simultaneous activation of opening and closing ’78) than Amia. muscles by a preprogrammed neural program One salient feature of the suction feeding at roughly equivalent frequencies and recruit- mechanism is the increasing overlap in an- ment rates. The movement pattern (i.e., jaw- tagonistic muscle complexes with increasing opening followed by closing) depends, however, rapidity of the strike. In Polypterus, no overlap on the mechanical characteristics of the occurs between activity in the jaw opening peripheral jaw structures-i.e., the length- muscles of the expansive phase (sterno- tension curves for the opening and closing mus- hyoideus, hypaxial, and epaxial muscles) and cles, mechanical advantage of the opening and the jaw closing muscles (adductor mandibulae closing lever systems, the length of the muscles part 2, “temporalis”, and “pterygoideus”). at initial activation, and viscous and elastic Lepisosteus and Amia both show extensive properties of the joints and ligaments. This overlap in muscle activity and at least one com- hypothesis has been advanced for the control of ponent of the adductor mandibulae complex is tongue function in the salamanderBolitoglossa active simultaneously with jaw-opening mus- (Thextonet al., ’77) in which the protrusive and cles. Although the electromyograms of expan- retrusive muscles are simultaneously acti- sive and compressive muscles may occur simul- vated, but the actual movement pattern of the taneously, the forces generated by jaw opening tongue appears to be governed by differences in and closing muscles probably peak at different the initial length of the muscles relative to times to generate the observed pattern of bone their length-tension curves. movement (jaw opening followed by closing). The latter hypothesis involves the initial input of a preprogrammed oscillator, but obvi- Neural control of the strike ates the need for precise preprogramming and The rapidity of the strike in Polypterus, rapid modification of recruitment rates and Lepisosteus, andAmia and the relatively invar- frequencies within an extremely short time iant nature of the movement pattern and elec- span (20 msec in Lepisosteus). In the gar, both tromyographic profile suggests that the initial the sternohyoideus and adductor mandibulae strike is governed by a preprogrammed output could be activated simultaneously. from the central nervous system that is not If the sternohyoideus is near its optimum subject to peripheral feedback. In contrast to length for force generation (control of pectoral observations on piscivorous cichlid fishes girdle position would be important in regulat- (Liem, ’781, no versatility was detected in the ing this parameter) and/or the mechanical ad- preprogrammed patterns elicited in response to vantage of the sternohyoideus is greater than different prey items. For example, strikes on that of the adductor mandibulae, mouth open- prey completely immobilized by anesthesia oc- ing will result. As the mouth opens, the sterno- curred just as rapidly and with an identical hyoideus will not remain at the same position pattern of muscle activation to those involving on the length tension curve and the mechanical 310 GEORGE V. LAUDER, JR. advantage may decline, thus allowing the Neurocranial elevation also plays snimportant adductor mandibulae (which may now have role during the expansive phase to increase the been stretched to near its optimum length for gape both in the fishes examined here and in force generation) to close the mouth. Testing of teleosts (e.g., Saluelinus, Lauder and Liem, these two alternative hypotheses awaits accu- '80; Hoplias, Lauder, '79; Lepomis, Lauder rate determination of length-tension curves and Lanyon, in press; Perca, Osse, '69; cichlids, and the mechanics of fish jaw muscles. Liem and Osse, '75). Thus, lifting of the head during the strike appears to be a fundamental The primitive actinopterygian feeding characteristic of the feeding mechanism and mechanism: Jaw mechanics has been retained throughout the radiation of Two key features of the mouth opening the actinopterygian fishes. mechanism in Polypterus, Lepisosteus, and Mandibular depression during the strike in Amia are hypothesized to be primitive for ac- palaeoniscoid fishes probably was accom- tinopterygian fishes: 1) neurocranial elevation plished by contraction of the obliquus inferi- during the expansive phase by the epaxial oris, sternohyoideus, and possibly the bran- muscles, and 2) mandibular depression medi- chiomandibularis. The occurrence of a man- ated by contraction of the sternohyoideus mus- dibulohyoid ligament is hypothesized to be cle via the mandibulohyoid ligament. primitive not only for the Actinopterygii, but Schaeffer and Rosen ('61) noted that eleva- also for the Teleostomi as a whole (Table 11, and tion of the neurocranium seems to have been an in palaeoniscoidsthis ligament served to trans- important mechanism acting to increase the fer the posterodorsal hyoid movement to the gape during feeding in palaeoniscoid fishes. mandible. Schaeffer and Rosen ('61:188) pro-

Fig. 18. Diagrammatic model of the feeding mechanism in a primitive actinopterygian fish (palaeoniscold) Note the anteroventrally inclined axis of the suspensorium, the three divisions of the adductor mandibulae (see text) which are enclosed in a maxillary-palatoquadrate chamber, the presence of a mandibulohyoid ligament that transmits the force of sternohyoideus contraction to the retroarticular process ofthe lowerjaw, the adductor opercuh which is continuous anteriorly with the adductor hyomandibulae (dashed line on suspensorium), and the presence of intermandibularis posterior and interhyoideus muscles in the intermandibular region. Fibers of these muscles do not insert onto the gular plate (see text). TABLE 1. The distribution of certain anatomical and functional attributes ofprimitcuefishes Greatest level of generality at which character is hypothesized Known character distribution: Character to be synapomorphic reference

1) Mandibular depression mediated by the Teleostomi’ Actinistia: Lauder (‘80) sternohyoideus via the hyoid apparatus Dipnoi: Lauder (unpublished data) and mandibulohyoid ligament. Actinopterygii: this paper 2) Double burst pattern of activity during Actinopterygii Actinopterygii: this paper the strike in the adductor operculi. 3) Addudor operculi continuous with adductor Actinopterygii Actinopterygii: this paper hyomandibulae. Lack of adductor arcus palatini. 4) Lack of a preparatory phase in the strike. Actinopterygii Actinopterygii: this paper 5) Stemohyoideus divided into two halves by Teleostomi Actinopterygii: this paper a median septum or completely separated. Actinistia: Millot and Anthony (’58) Dipnoi: Lauder (unpublished data) 6) Sternohyoideus not divided into two Teleostei Teleostei: Winterbottom (‘74) halves by a median septum. 7) Intermandibular musculature consisting of Teleostomi Actinistia: Millot and Anthony (‘58) intermandibularis posterior and interhyoideus. Actinopterygii: this paper Choanata: Jarvik (’63) 8) Anterior division of the adductor mandibulae Elasmobranchiomorphi Elasmobranchiomorphi: Lightoller (‘39) complex extending posteroventrally from the plus Actinopterygii: this paper suborbital region toinsert on lower jaw. Teleostomi “Rhipidistia”: Thomson (‘67) Acanthodii: Lauder (unpublished) 9) Presence of a levator onerculi couolinnLy Halecostomi Halecomomhi: this DaDerL1 mediating mandibular depression. Teleostei: iiem (‘70) 10) Maxilla mobile and free from the cheek. Halecostomi Halecostomi: Patterson and &sen (‘77)

‘Teleostomi = Acanthodii + Actinopterygii + Dipnoi + Actinistia + Choanata 312 GEORGE V. LAUDER, JR. posed that the mandible was depressed by the not function to depress the mandible but rather geniohyoid muscles. Palaeoniscoids lacked a to compress the buccal cavity during chewing geniohyoideus because the primitive condition following prey capture (interhyoideus), or to of the actinopterygian intermandibular muscu- adduct the mandibular rami and suspensorium lature involves the presence of an interman- during the expansive phase (intermandibularis dibularis posterior and an interhyoideus dorsal posterior). to the gular elements (Table 1; Fig. 18). These Movement of the operculum and suboper- muscles may have been active during mouth culum was controlled by the dilator and opening (as in Amia) but could not have con- adductor operculi which permitted limited con- tributed to mandibular depression because of trol over opercular movements during feeding. their line of action. Adduction of the operculum occurred in a bi- The mechanism of mandibular depression phasic pattern during the strike, the first phase proposed here as primitive for actinopterygians occurring at the start of buccal expansion and also may be primitive at a much greater level of the final phase following closure of the jaws. generality. Recently, I have reconsidered the feeding mechanism in the living coelacanth, Muscle homologies and comparative anatomy Latimeria, (Lauder, '80) and proposed that The homologies of the divisions of the os- mandibular depression is mediated by the teichthyian adductor mandiblae complex have sternohyoideus and ligamentous connections not been considered since the earlier investiga- between the hyoid and mandible. All living tions of Edgeworth ('351, Souche ('321, Lightol- lungfkhes have mandibulohyoid ligaments. ler ('39), Luther ('13), and to some extent Allis Although the structure of the hyoid arch is (1897). poorly known in acanthodian fishes, the occur- The adductor mandibulae complex of primi- rence of a small "accessory element" (= inter- tive living actinopterygians may be divided hyal?) between the distal end of the hyoman- into three separate divisions: 1)an anterior dibula and the ceratohyal (Miles, '73) suggests (suborbital) division originating on the an- that hyoid retraction may have played a key terior of the palatoquadrate and extending role in mediating mandibular depression. This posteroventrally to insert on the lowerjaw; 2) a role of the sternohyoideus and hyoid apparatus medial division that runs nearly dorsoven- in the feeding mechanism thus may be primi- trally from the palatoquadrate to the mandible; tive for the Teleostomi (Table 1). and 3) a posterolateral division originating A comparative analysis of the structure of from the posterior aspect of the palatoquadrate the jaw in primitive living actinopterygians, (see Fig. 18). This partition is based on the Latimeria, and lungfkhes, suggests the follow- topography of the adductor complex in the ing model for the feeding mechanism of adult as well as on the embryology of the jaw palaeoniscoid fishes (Fig. 18): The suspensory adductor complex (Edgeworth, '35).The three apparatus possessed rather limited lateral mo- adductor divisions in primitive actinoptery- bility and the axis of rotation during mediolat- gians represent a separation into discrete com- era1 excursion was anteroventrally inclined ponents of the primitively large fan-shaped (Fig. 18).Epaxial musculature inserting on the adductor mandibulae of gnathostomes. neurocranium was instrumental in increasing Within this scheme, the levator maxillae su- the gape during feeding, and lateral head perioris ofdmia, so-named by Allis (1897) be- movements were probably effected by the ob- cause of its supposed homology with the levator liquus superioris (Fig. 18: OBS) division of the maxillae superioris of selachians, is homolo- hypaxial muscles. The pectoral girdle re- gous to the suborbitalis of selachians (= levator mained nearly stationary during feeding as a labii superioris of Luther, '13; muscle preor- result of antagonistic activity in the obliquus bitaire of Souche, '32) as an anterior division of inferioris and sternohyoideus muscles, which the adductor mandibulae. Thus, both the sel- controlled hyoid retraction and, indirectly, achian suborbitalis and the levator maxillae mandibular depression. The interhyal bone superioris of Amia are derivatives of the an- (Gardiner, '73) functioned to decouple move- terior fibers of the primitive fan-shaped os- ments of the hyoid from those of the suspen- teichthyian adductor mandibulae. sorium and increased the mechanical advan- In Lepisosteus, the palatomandibularis tage of hyoid retraction. The gular elements minor and major (Luther, '13) correspond to the and branchiostegal rays were relatively immo- anterior adductor division (Table 21, whereas bile and the intermandibular musculature did Polypterus lacks an anterior adductor compo- ACTINOPTERYGIAN FEEDING MECHANISMS 313

nent. The homologies of the medial and posterolateral adductor components are indicated in Table 2. The name "levator maxillae superioris" (LMS) was used by Rosen and Patterson ('69)to refer to certain divisions of the adductor mandibulae in paracanthopterygian fishes; Winterbottom ('74) pointed out that this muscle is not homologous to parts three and four of the LMS in Amia. Despite this, Nelson ('74) and Casinos ('78) continued touse"1evator maxillae superioris" for paracanthopterygian muscles. The Osteoglossomorpha, Elopomorpha, Clupeomorpha, and basal euteleosteans lack any muscle comparable in origin, insertion, and ontogeny to the LMS ofAmia. This renders any interpretation of paracanthopterygian muscles as LMS homologues highly unpar- simonius. The anterior division of the primitive osteichthyian adductor complex to which the LMS ofAmia is homologous (Table 2) has been lost in all teleost fishes. The intermandibularis musculature of primitive, living actinopterygians consists of an intermandibularis posterior spanning the mandibular rami, and an interhyoideus muscle, extending anteriorly from the ceratohyal to insert dorsal to the gular plate (Figs. 1-3: IMP, IH). This condition is apparently primitive not only for the Actinopterygii (Fig. 181, but also for the Teleostomi (Table 11, because Latimeria (Millot and Anthony, '58; pers obs.) and choan- ates (Jarvik, '63) have intermandibular musculature nearly identical to that of actinopterygians. In none of these groups do the fibers of the intermandibularis posterior or interhyoideus actually insert on the gular plate (where this exists). Two groups of teleosts, the Osteoglossomorpha and Elopomorpha, retain identifiable intermandibularis and interhyoideus components of the geniohyoideus (Greenwood, '71; Liem, '67a). The presence of an intermandibularis anterior is an additional synapomorphy between Amia and teleosts. The anatomy of the sternohyoideus muscle has been the subject of a number of recent in- vestigations. Winterbottom ('74) and Green- wood ('77) commented on the difficulty of pre- cisely defining the posterior limit of this muscle, because in teleosts, its posterior fibers are often continuous with those of the anterior obliquus inferioris. Commonly, teleosts have two transverse tendinous inscriptions in the sternohyoideus, and Winterbottom ('74) proposed that the posterior limit of this muscle be set, somewhat arbitrarily, at the third myosept. 3 14 GEORGE V. LAUDER, JR. The comparative anatomy of the sternohy- ive) major osteological differences between the oideus in Polypterus, Lepisosteus, and Amia skulls of Lepisosteus, Polypterus, and Amia. suggests that the primitive actinopterygian The major variations in cranial myology occur sternohyoideus was paired (divided completely in the adductor mandibulae complex, whereas into right and left halves), originated from the all other jaw musculature is constant in posi- anteroventral face of the pectoral girdle and tion. had no fiber continuity with the obliquus infe- The origin and evolution of rioris, and was divided transversely by two myosepts. Winterbottom’s (’74)proposal, thus, inertial suction feeding has some justification because the position of The inertial suction strategy of prey capture the third myosept may be expected to approxi- involves the creation of a low pressure center in mate the primitive attachment of the sterno- the buccal cavity by rapid mouth opening and hyoideus to the pectoral girdle. expansion of the buccal floor. The pressure dif- The paired condition of the sternohyoideus ferential between the buccal and opercular may be primitive for the Teleostomi (Table 1) cavities and the surrounding water results in a because Latimeria and the living lungfishes- flow of water into the mouth. If the velocity of also have a divided sternohyoideus. In teleost flow at the position of the prey is sufficiently fishes, the sternohyoideusis consolidated into a great, the prey will be carried into the mouth single median muscle, although Heterotis still with the flow of water. Forward body velocity of has a remnant of the primitive longitudinal the predator may also be used in prey capture to sternohyoideus septum (Greenwood, ’7 152). varying degrees, some fishes relying exclu- The myology of primitive actinopterygians sively on pursuit, whereas others remain es- has remained remarkably conservative and sentially stationary and capture prey by suc- relatively few differences in cephalic myology tion feeding. occur despite numerous (potentially disrupt- The features of the actinopterygian feeding

PRIMITIVE ACTINOPTERYGIAN HALECOSTOME

-41 NEUROCRANIUM I NEUROCRANIUM I-- %\ *I r r 1 m\ /MOUTH

r OPERCULUM

L - -1 INTEROPERCULUM ACTINOF’TERYGIAN FEEDING MECHANISMS 315 mechanism, identified here as morphological Redfieldiiform fishes also lack the mor- correlates of an inertial suction feeding strat- phological features identified here as correlates egy of prey capture, complement the characters of a suction-feeding mechanism. The maxilla is used by Patterson and Rosen (‘77) in defining fixed to the cheek, an interopercular bone is the halecostome fishes (= Halecomorphi plus absent, and a single mechanism exists for man- Teleosteil-i.e., the free maxilla and inter- dibular depression. operculum. The occurrence of a maxilla that is The model of the palaeoniscoid feeding no longer immovably attached to the dermal mechanism presented above suggests that red- cheek bones is a major feature of a suction feed- fieldiiform fishes used the primitive actino- ing mechanism because maxillary swing pre- pterygian mechanism of mandibular depres- vents water inflow from the sides of the mouth sion mediated by the sternohyoideus via the and results in more anteroposteriorly oriented mandibulohyoid ligament; the negative pres- streamlines. sures within the buccal cavity were probably In addition, the halecostome fishes are the small (about 50 cm H,O). first to show a levator operculi-opercular series-mandible coupling that mediates man- ACKNOWLEDGMENTS dibular depression (Table 1). The inter- This study was supported by an NIH Predoc- opercular bone is a key element for transmit- toral Fellowship in Musculoskeletal Biology ting the pull of the levator operculi to the (51‘32 GM077 117). mandible and this bone is absent in all Many people provided assistance in the nonhalecostome actinopterygians. The pres- course of this study. Natalie Bigelow, Debbie ence of a levator operculi coupling in addition to Sponder, Steve Norton, Charlie Gougeon, the primitive mechanism of mandibular de- Larry Nobrega, Hans Dieter-Sues, and espe- pression (the sternohyoideus-hyoid appara- cially David Kraus aided with the experi- tus-mandible coupling) allows the two cou- mental work. Karsten Hartel, Stan Weitzman, plings either to act synergistically to effect and Donn &sen kindly provided specimens for rapid mandibular depression or to function at dissection. Bill Fink, William Hylander, R. different stages of the expansive phase. The Winterbottom, E.O. Wiley, P.H. Greenwood, presence of two independent mechanisms con- and Karel F. Liem provided many helpful trolling mandibular depression (Fig. 19) is a comments on the manuscript. I especially shared feature of the halecostome fishes as is thank Karel Liem for extensive discussion and the inertial suction strategy of prey capture. support throughout this study. Verraes (’77) has shown in Salmo that the mandibulohyoid ligament is present at hatch- LITERATURE CITED ing, while the interoperculomandibular liga- Alexander, R. McN. (1966)The functions and mechanisms of theprotrusible upperjaws of two species ofcyprinid fish. J. ment develops during the late larval phase. The 2001. (Lond.),249:28%296. former is thus a more general character (sensu Alexander, R. McN. (1967)The functions and mechanisms of Nelson, ’78) than the latter. the protrusible upper jaws of some acanthopterygianfish. Hutchinson (’73) suggested that a group of J. Zool. (Land.), 15Zr4g64. Alexander,R. McN. (1969)Mechanics ofthe feeding action of “chondrosteans,” the redfieldiiform fishes, a cyprinid fish. J. Zool. (Lond.),159:l-15. were suction feeders. This conclusion was based Alexander, R. McN. (1970)Mechanics of the feeding action of on the suggestion that the buccal cavity could various teleost fishes. J. 2001. (Lond.),262:14&156. be expanded by posteroventral movement of Allis, E.P. (1897) The cranial muscles and cranial and first spinal nerves in Amia calm. J. Morphol., 12:487-772. the pectoral girdle and that ‘f. . . considerable Allis, E.P. (1919a) The homologies of the maxillary and negative pressure in the buccal cavity could be vomer bones of Polypterus. Am. J. Anat., 25:349-394. achieved by independent abduction of the Allis, E.P. (1919b)The innervation of the intermandibularis branchial basket.” Abduction (ventral rota- and geniohyoideusmuscles of the bony fishes. Anat. Rec., 16:295-307. tion) of the branchial basket can be achieved Allis. E.P. (1920) On certain features of the otic region of the independently of mouth opening in all fishes, chondroeranium of Lepidosteus, and compa&on with and is not a correlate of large negative pres- other fishes and higher vertebrates. Proc. Zool. SOC. sures in the buccal cavity. Furthermore, (Lond.),1920t245-266. Allis, E.P. (1922) The cranial anatomy of Polypterus, with posteroventral movement of the pectoral girdle special reference to Polypterus bichir. J. Anat., 56:18& will contribute little to a reduction of the pres- 294. sure within the mouth cavity due to the small Ballintijn, C.M., A. van den Berg, and B.P. Egberink (1972) volume change. Hutchinson’s observation that An electromyographic study of the adductor mandibulae complex of a free-swimming carp (Cyprinuscarpio L.) dur- the sternohyoideus may have been a large mus- ing feeding. J. Exp. Biol., 57.261-283. cle suggests that the pectoral girdle might ac- Basmajian, J.V., and G. Stecko (1962) A new bipolar elec- tually have been protracted during feeding. trode for electromyography. J. Appl. Physiol., 27:849. 316 GEORGE V. LAUDER, JR.

Casinos, A. (1978) The comparative feeding mechanism of during feeding in the gar, Lepisosteus oculatus. J. Exp. Gadidae and Macrouridae. I. Functional morphology of the Biol. (in press). feeding apparatus. Gegen. Morph. Jahrb., 124,436449. Liem, K.F. (1967a) A morphological study of Luciocephalus Dobben, W.H. van (1935) iiber den Kiefermechanismus der pulcher, with notes on gular elements in other recent tele- Knochenfische. Arch. Neerl. Zool., 2: 1-72. osts. J. Morphol., 121:10%134. Dutta, H.M. (1968) Functional Morphology of the Head of Liem, K.F. (1967b)Functional morphology of the head of the Anabas testudineus (Block). Ph. D. Thesis, Univ. Leiden, anabantoid teleost fish, Helostoma ternrnincki. J. Mor- 146 pp. phol., 121: 13t5158. Edgeworth, F.H. (1935)The Cranial Muscles of Vertebrates. Liem, K.F. (1970) Comparative functional anatomy of the Cambridge University Press, London. 493 pp. Nandidae (Pisces: Teleostei). Fieldiana Zool., 56: 1-166. Elshoud-Oldenhave, M.J.W., and J.W.M. Osse (1976) Func- Liem, K.F. (1973) Evolutionary strategies and morphologi- tional morphology of the feeding system in the ru%- cal innovations: cichlid pharyngeal jaws. Syst. 2001. Gyrnnocephnlus cernua (L. 1758t(Teleostei,Percidae). J. 22t426441. Morphol., 150:39%422. Liem, K.F. (1978) Modulatory multiplicity in the functional Gans, C. (1969) Functional components versus mechanical repertoire of the feeding mechanism in cichlid fishes. I. units in descriptive morphology. J. Morphol., 128:365- Piscivores. J. Morphol., 158:323-360. 368. Liem, K.F., and G.V. Lauder The evolution and inter- Gardiner, B.G. (1973) Interrelationships of teleostomes. In: relationships of the actinopterygian fishes. In: Fish Interrelationships of Fishes. P.H. Greenwood, R.S. Miles, Neurobiology and Behavior. University of Michigan Press. and C. Patterson, eds. Academic Press, London, pp. 105- (in press) 135. Liem, K.F., and J.W.M. Osse (1975) Biological versatility, Gardiner, B.G., and A.W.H. Bartram (1977) The homologies evolution, and food resource exploitation in Africancichlid of the ventral cranial fissures in osteichthyians. In: fishes. Am. Zool., 15:427-454. Problems in Vertebrate Evolution. S.M. Andrews, R.S. Lightoller, G.H.S. (1939) Probable homologues. A study of Miles, and A.D. Walker, eds. Academic Press, London, pp. the comparative anatomy of the mandibular and hyoid 227-245. arches and their musculature. Trans. R. Soc. (London), Greenwood, P.H. (1971) Hyoid and ventral gill arch muscu- 24: 34%445. lature in osteoglossomorph fishes. Bull. Br. Mus. Nat. Luther, A. (1913) herdie von N. trigeminus versgorts Mus- Hist. Zool., 22:l-55. kulatur der Ganicden und Dipneusten. Acta SOC.Scient. Greenwood,P.H. (1977)Notes on the anatomy and classifica- Fern., 4l:l-72. tion of elopomorph fishes. Bull. Br. Mus. Nat. Hist. Zool., Mayhew, R.L. (1924) The skull ofLepidosteuspZatostornus.J. 32:6&102. Morphol., 38:315341. Greenwood, P.H., R.S. Miles, and C. Patterson (1973)Inter- Miles, R.S. (1973) Relationships of acanthodians. In: Inter- relationships of Fishes. Academic Press, London. relationships of Fishes. P.H. Greenwood,R.S. Miles, and C. Gunther, K., and K. Deckert (1953) Morphologisch- Patterson, eds. Academic Press, London, pp. 63-104. anatomische und vergleichendokologische Unter- Millot, J., and J. Anthony (1958) Anatomie de Latimeria suchungen iiber die Leistungen des Viscerocranial- chalurnnne. I. Squelette, muscles et formations de sou- Apparates bei Tiefseefischen der Gattung Cyclothone tiens. Centre National de la Recherche Scientifique, Paris. (Teleostei, Isospondyli). 2. Morph. Okol. Tiere., 42:l-66. Nelson, G.J. (1969)Gill arches and the phylogeny of fishes, Holmquist, 0. (1910) Der Musculus protractor hyoidei und with notes on the classification of vertebrates. Bull. Am. der Senkungsmechanismus des Unterkiefers bei den Mus. Nat. Hist., 141:475552. Knochenfkchen. Acta Univ. Lund., 6: 1-27. Nelson, G.J. (1974) Paracanthopterygii. In: Encyclopedia Hutchinson, P. (1973) A revision of the redfieldiiform and Britanica, 15th. perleidiform fishes from the Triassic of Bekker's Kraal Nelson, G.J. (1978) Ontogeny, phylogeny, paleontology, and (South Africa) and Brookvale (New South Wales). Bull. Br. the biogenetic law. Syst. Zool., 27.324-345. Mus. Nat. Hist. Geol., 22:233-354. Nyberg, D.W. (1971) Prey capture in the largemouth bass. Jarvik, E. (1963) The composition of the intermandibular Am. Midl. Nat., 86:128-144. division ofthe head in fish and tetrapodsand the diphyletic Osse, J.W.M. (1969) Functional morphology of the head of origin of the tetrapod tongue. Kungl. Sven. Veten. Handl., the perch (Percafluuiatilis L): an electromyographic study. 9:l-74. Neth. J. Zool., 19;28%392. Kampf, W. (1961) Vergleichende funktionsanalytische Un- Osse, J.W.M. (1976) Mechanismes de la respiration et de le tersuchungen an den Viscerocranien einiger rauberisch prisedesproies chezAmia calm Linnaeus. Rev. Trav. Inst. lebender Knochenfische. Zool. Beitr. (N.Ser.), 6:391-496. Pesches Marit., 40:701-702. Kirchhoff, H. (1958) Funktionell-anatomische Unter- Patterson, C. (1973)Interrelationships of holosteans. In: In- suchung des Visceralapparates von Clupea harengus L. terrelationships of Fishes. P.H. Greenwood, R.S. Miles, 2001. Jahrb. Anat., 76:461-540. and C. Patterson, eds. Academic Press, London, pp. 23% Lauder, G.V. (1979) Feeding mechanisms in primitive tele- 305. osts and in the halecomorph fish Amia calua. J. Zool., Patterson, C., and D.E. Rosen (1977) Review of ichthyodecti- (Lond.), 187:54%578. form and other Mesozoic teleost fishes and the theory and Lauder, G.V. (1980) The role of the hyoid apparatus in the practice of classifying fossils. Bull. Am. Mus. Nat. Hist., feeding mechanism of the coelacanth Latimeria chalum- 158:81-172. nae. Copeia ltl-9. Rosen, D.E., and C. Patterson (1969) The structure and re- Lauder, G.V., and L.E. Lanyon Functional anatomy of feed- lationships of the paracanthopterygian fishes. Bull. Am. ing in the bluegill sunfish, Lepornis macrochirus: In vivo Mus. Nat. Hist., 141:357-474. measurement of bone strain. J. Exp. Biol. (in press) Schaeffer, B., and D.E. Rosen (1961) Major adaptive levels in Lauder, G.V., and K.F. Liem (1980) The feeding mechanism the evolution of the actinopterygian feeding mechanism. and cephalic myology of Saluelinus fontinulis: form, func- Amer. Zool., 1:187-204. tion, and evolutionary significance. In: Charrs: Salmonid Souche, G. (1932) Morphologie comparative des muscles ele- Fishes of the GenusSaZuelinus. E.K. Balon, ed. Junk Pub- vateurs de la mandibule chez les poissons. Mem. SOC.Sci- lishers. The Netherlands. ChaDt. 10. I. ent. Phv. Nat. de Bordeaux. 7e Serie. 3;l-292. Lauder, G.V., and S.F. Norton Asymmetrical muscle activity Tchernavin, V.V. (1953) The feeding mechanisms of a deep ACTINOP'I'ERYGIAN FEEDING MECHANISMS 317

sea fish Chauliodus sloani SJmeider. British Museum, gairdneri Richardson, 1836 (Teleostei: Salmonidae). J. N.H., London. Morphol. ,151 r 11 1- 120. Thexton, A.J., D.B. Wake, and M.H. Wake (1977) Tongue Wiley, E.O. (1976) The phylogeny and biogeography of fossil function in the salamander Bolitoglossa occidentalis. and recent gars (Actinopterygii: Lepisosteidae). Univ. Archs Oral Biol., 22r361-366. Kansas Mus. Nat. Hist. Misc. F'ubl., 64:l-111. Thomson, K.S. (1967) Mechanisms of intracranial kinetics in Wiley, E.O. (1979) Ventral gill arch muscles and the inter- fossil rhipidistian fishes (Crossopterygii) and their rela- relationships of gnathostomes, with a new classification of tives. J. Linn. SOC.Zool., 178:23%253. the Vertebrata. J. Linn. Soc., 67r149179. Verraes, W. (1977) Postembryonic ontogeny and functional Winterbottom, R. (1974) A descriptive synonymy of the anatomy of the ligamentum mandibulo-hyoideum and the striated muscles of the Teleostei. Proc. Acad. Nat. Sci. ligamentum interoperculo-mandibulare, with notes on the (Phil.), 125:225317. opercular bones and some other cranial elements inSalmo