q 2002 by the American Society of Ichthyologists and Herpetologists

Copeia, 2002(1), pp. 24–38

Kinematic Analysis of Suction Feeding in the Nurse Shark,Ginglymostoma cirratum (Orectolobiformes, Ginglymostomatidae)

PHILIP J. MOTTA, ROBERT E. HUETER, TIMOTHY C. TRICAS, AND ADAM P. SUMMERS

Inertial suction feeding is known to occur in some sharks, but the sequence andtemporal kinematics of head and jaw movements have not been defined. We inves-tigated the feeding kinematics of a suction feeding shark, the nurse shark Gingly-mostoma cirratum, to test for differences in the timing and magnitude of feedingcomponents with other shark taxa when sharks were fed pieces of bony fish. Thir-teen kinematic variables were measured from high-speed video recordings. Foodcapture in this species consists of expansive, compressive, and recovery phases, asin most other sharks, but there is little or no cranial elevation. Mean time to maxi-mum gape (32 msec) is the fastest recorded for an elasmobranch fish. Other rela-tively rapid events include mandibular depression (26 msec), elevation (66 msec),and total bite time (100 msec). Buccal valves assist the unidirectional flow of waterinto the mouth and out of the gill chambers. Food capture under these experimentalconditions appears to be a stereotyped modal action pattern but with significantinterindividual variability in timing of kinematic events. Ginglymostoma cirratum ex-hibits a suite of specializations for inertial suction feeding that include (1) the for-mation of a small, anteriorly directed mouth that is approximately round and lat-erally enclosed by modified labial cartilages; (2) small teeth; (3) buccal valves toprevent the backflow of water; and (4) extremely rapid buccal expansion. Sharksthat capture food by inertial suction have faster and more stereotyped capture be-havior than sharks that primarily ram feed. Inertial suction feeding, which hasevolved multiple times in sharks, represents an example of functional convergencewith inertial suction feeding bony fishes.

THE feeding apparatus of vertebrates is acomplex mechanical system with obviousimportance to fitness. Understanding the dy-namics of the prey capture mechanism in anevolutionary framework gives insight into theconstraints and flexibility that govern changesin this key system (Lauder, 1990). Studies of car-tilaginous fishes, bony fishes, amphibians, andreptiles have produced evolutionary hypothesesregarding the transition from aquatic to terres-trial prey capture (Lauder, 1985; Lauder andReilly, 1994; Wilga et al., 2000). Sharks, skates,and rays (elasmobranchs) possess a radically dif-ferent suspensorium and jaw structure fromthat of the bony fishes. In spite of the structuraldifferences, elasmobranchs and bony fisheshave evolved a similar suite of prey capturemechanisms including suction, grasping, biting,gouging, and filter feeding (Moss, 1977; Mottaand Wilga, 2001).

Aquatic prey capture in sharks and bony fish-es may be accomplished through ram, biting, orsuction. Ram feeding, which occurs when thepredator overswims the prey, may encompassfeeding events that establish a continuous cur-rent in the mouth and over the gills. This occursin the filter feeding basking shark Cetorhinusmaximus or when the predator approaches the

prey with an open mouth resulting in some an-terior or lateral displacement of water, such asin the white shark Carcharodon carcharias (Dia-mond, 1985; Tricas and McCosker, 1984; Sims,2000). Biting, which may accompany ram, oc-curs when a shark swims toward the prey, stops,and then bites or removes pieces off the prey,such as in the cookiecutter shark Isistius brasi-liensis (LeBoeuf et al., 1987; Motta and Wilga,2001).

The mechanics and function of suction feed-ing are controversial (Muller et al., 1982; Nor-ton and Brainerd, 1993; Van Damme and Aerts,1997). The clearest terminology is that of VanDamme and Aerts (1997), who suggest that suc-tion feeding includes a continuum of behaviorsranging from sucking in just enough water tokeep the prey item from escaping as the pred-ator swims over it (compensatory suction) tothe entrainment of the prey item in a columnof water that is transported toward the motion-less predator (inertial suction). Bony fishes spe-cialized for inertial suction feeding typicallyhave a small, laterally enclosed mouth, protru-sible upper jaw, reduced dentition, and stronglydeveloped abductor muscles. Buccal expansionis large and rapid, and there is a wave of expan-sion and subsequent compression that travels

25MOTTA ET AL.—NURSE SHARK FEEDING

Fig. 1. Phylogeny of neoselachians according toShirai (1996) with placement of Carcharhinus perezi ac-cording to G. J. P. Naylor (pers. comm.) incorporatingprey capture types (R 5 ram feeding; S 5 suctionfeeding; S/F 5 suction filter feeding) based on thefollowing kinematic studies: Heterodontus francisci 5Edmonds et al. (2001); Ginglymostoma cirratum 5 thisstudy and Robinson and Motta (2002); Orectolobus ma-culatus 5 Wu (1994); Rhincodon typus 5 Clark and Nel-son (1997); Carcharodon carcharias 5 Tricas andMcCosker (1984); Cephalosyllium ventriosum 5 Ferry-Graham (1997); Triakis semifasciata 5 Ferry-Graham(1998); Carcharhinus perezi 5 Motta and Wilga (2001);Negaprion brevirostris 5 Motta et al. (1997); Sphyrna ti-buro 5 Wilga and Motta (2000); Notorynchus cepedianus5 personal observation (PJM); Squalus acanthias 5Wilga and Motta (1998a); Rhinobatos lentiginosus 5Wilga and Motta (1998b).

posteriorly such that the engulfed body of waterand prey is carried into and through the buccalcavity (Lauder, 1980a; Muller et al., 1982; Liem,1993). However, suction feeding in sharks andrays has only been investigated in a few species(Wu, 1994; Ferry-Graham, 1998; Pretlow-Ed-monds, 1999). Buccal expansion is rapid andincludes the use of labial cartilages and upperjaw protrusion to form a somewhat round andlaterally enclosed mouth that is small to inter-mediate in size, and an anterior to posteriorwave of expansion and compression draws thefood into the mouth (Ferry-Graham, 1998; Wil-ga and Motta, 1998a).

Inertial suction feeding is hypothesized to bethe ancestral form of prey capture for the bonyfishes (Lauder, 1985). In contrast, fossil evi-dence and observations of morphologically sim-ilar, but extant sharks indicate that the ancestralprey capture behavior in elasmobranchs prob-ably involved grasping the prey whole or tearingpieces from it, with minimal upper jaw protru-sion or suction. Many of these Paleozoic sharkshad a grasping dentition, long slitlike jaw andamphistylic jaw suspension that permitted littleupper jaw kinesis (Schaeffer, 1967; Moy-Thomasand Miles, 1971; Maisey, 1980). A variety of ex-tant elasmobranchs use inertial suction to somedegree as their primary feeding method [spinydogfish Squalus acanthias (Wilga and Motta,1998a); leopard shark Triakis semifasciata (Russo,1975; Talent, 1976; Ferry-Graham, 1998); wob-begong Orectolobus maculatus, nurse shark G. cir-ratum, and whale shark Rhincodon typus (Wu,1994; Clark and Nelson, 1997; Robinson andMotta, 2002); horn shark Heterodontus francisci(Strong, 1989; Pretlow-Edmonds, 1999; Ed-monds et al., 2001); guitarfish Rhinobatos lenti-ginosus (Wilga and Motta, 1998b); and perhapsthe angel shark Squatina californica (Fouts andNelson, 1999)]. Inertial suction feeding elas-mobranchs are found in at least eight families,often nested within clades that contain ram andcompensatory suction feeders, indicating thatinertial suction feeding has evolved indepen-dently in several elasmobranch lineages (Fig. 1)(Motta and Wilga, 2001). However, functionalconvergence of inertial suction feeding insharks has received little attention (Moss, 1977;Motta and Wilga, 1999).

Orectolobiformes, which include the nursesharks and wobbegongs, are considered special-ized inertial suction feeders based on their anat-omy and feeding kinematics (Moss, 1977; Wu,1994; Motta and Wilga, 1999). Ginglymostoma cir-ratum is common to shallow waters throughoutthe West Indies and south Florida, where itfeeds on small fishes and crustaceans (Compag-

no, 1984; Castro, 2000). This benthic shark rou-tinely uses inertial suction, generating pressuresas low as2760 mm Hg, to capture its prey (Ta-naka, 1973; Robinson and Motta, 2002).

During prey capture many fishes demonstrateinherent intra- and interindividual variability intheir kinematic and motor activity patterns, inaddition to the ability to modulate or alter thesepatterns in response to differing food types andpresentations (Liem, 1980a; Wainwright andTuringan, 1993; Nemeth, 1997). Variability mayoccur independently of the ability to modulateand may occur among feedings within a foodcategory (Liem, 1978; Chu, 1989, Ferry-Gra-ham, 1997). There is ample evidence of exten-sive variation of prey capture kinematics andmotor patterns within species of sharks and rays(Motta et al., 1997; Wilga and Motta, 1998a; Fer-ry-Graham, 1998), but it is unclear whether anapparently morphologically specialized obligatesuction feeding shark such as G. cirratum woulddemonstrate variability and what effect if anythis would have on prey capture ability.

In this study, we investigated the kinematicsof suction food capture in G. cirratum and com-pared it to food capture in other elasmo-branchs. Our goals were to (1) determine towhat extent G. cirratum relies on inertial suctionfeeding when presented with one food type; (2)

26 COPEIA, 2002, NO. 1

describe the kinematic pattern of suction foodcapture; (3) compare the kinematics of feedingto those of other cartilaginous and bony fishes;and (4) investigate interindividual variability infood capture kinematics and discuss this in lightof prey capture.

MATERIALS AND METHODS

Experimental procedure.—Specimens of juvenileG. cirratum were collected in Florida Bay northof the Florida Keys and held at Mote MarineLaboratory, Sarasota, Florida, in 5 m diametercircular holding tanks with natural seawater.Four males and one female, ranging from 64–98 cm total length (mean 5 78, SD 5 13 cm),were used in the experiments. These immaturesharks probably included age class 31 years andolder, although there are no published studieson G. cirratum aging prior to year 3 (Carrier andLuer, 1990; J. Castro, pers. comm.). An addi-tional 85 cm TL female was video recorded toprepare the composite photograph of prey cap-ture and to digitize the kinematic profile of aprey capture event. Approximately two weeksprior to the experiments, each animal was trans-ferred to a 2.4 m diameter, 1400-liter semicir-cular tank with a 0.5 3 1.7 m acrylic windowand fed cut pieces of fish three times a week.Water temperature in the experimental tankranged from 26–298 C. Pieces of fillets of Atlan-tic thread herring (Opisthonema oglinum) andcrevalle jack (Caranx hippos) were cut to approx-imately half the mouth diameter (mouth di-ameter ranged from 4–8 cm) and placed on thefloor of the tank or on a platform. Prey refershere to dietary items captured by natural feed-ing, and food refers to pieces of or whole fooditems offered under experimental conditions.Food size was approximately 2.5 3 3 3 1 cm formost bites. The animal was either conditionedto take the food from the floor of the tank infront of the window, or on a small plexiglassplatform raised approximately 25 cm off thetank floor. In the latter case the shark suckedthe food through a 2.5-cm hole in the clearplexiglass wall of the platform. In many cases,the shark remained stationary and propped onits pectoral fins as it took the food. Sharks be-came satiated after 22–44 bites, refusing furtherfood.

All experimental sharks were filmed duringfeeding with a high-speed video camera (NACHSV-200, 200 fps) or a Photosonics 1-PL camera(200 fps). A mirror placed at a 45–degree anglebelow the transparent floor of the tank provid-ed a simultaneous ventral view of the shark insome of the experiments. Illumination was pro-

vided with approximately 3000 watts of quartz-halogen lights. Because electromyographic ex-periments were being conducted simultaneous-ly on these sharks, they had bipolar fine wireelectrodes (0.06 mm diameter alloy wire) im-planted in their cranial muscles. Motta et al.(1997) determined that implantation of electro-myographic leads in the lemon shark Negaprionbrevirostris did not affect feeding kinematics.The 85-cm TL female G. cirratum used for thefood capture sequence and kinematic profile ofprey capture was videotaped at 250 fps with aRedlake PCI-1000 digital camera illuminated by300 watts of quart-halogen light.

Data collection.—Video images of the five exper-imental sharks were analyzed with a PanasonicAG-1970 video analyzer and monitor, and cinefilm with a LW Athena analysis projector. Onlyclear lateral images of capture bites were usedfor the analysis, and therefore 8–14 bites wereanalyzed per shark, totaling 53 bites in all. Thefollowing durations were determined for eachbite by counting the number of fields/frames(1 field/frame 5 5 msec) for each of the fol-lowing kinematic events: (1) mandibular de-pression 5 start of mandibular depression tomaximum mandibular depression; (2) mandib-ular elevation 5 maximum mandibular depres-sion to end of mandible elevation; (3) mandib-ular total 5 total time for mandibular depres-sion and elevation; (4) head onset 5 time frombeginning of mandibular depression to begin-ning of head elevation, if there was head ele-vation; (5) head elevation 5 start of head ele-vation to maximum head elevation; (6) headdepression 5 maximum head elevation to endof head depression; (7) head total 5 total timeof head elevation and depression; (8) labial on-set 5 time from start of mandibular depressionto start of labial cartilage protrusion; (9) labialprotrusion 5 start of labial protrusion to maxi-mum labial protrusion; (10) labial retraction 5maximum labial protrusion to complete labialretraction; (11) labial total 5 total time of labialprotrusion and retraction; (12) maximum gape5 start of mandibular depression to maximumgape; and (13) food enters mouth 5 start ofmandibular depression until the food enters themouth and is no longer visible. In lateral view,palatoquadrate protrusion was obscured by theprotrusion of the labial cartilages, but was visi-ble in more anterior views. Consequently the ki-nematics of palatoquadrate protrusion are onlyqualitatively described. Maximum hyoid depres-sion could not be accurately determined on themajority of bites because of the ventral bulging

27MOTTA ET AL.—NURSE SHARK FEEDING

TABLE 1. KINEMATIC VARIABLES RECORDED FOR FIVE Ginglymostoma cirratum AND RESULTS OF THE ONE-WAY AN-OVA OR KRUSKAL WALLIS AMONG SHARKS WITH KINEMATIC VARIABLES AS THE FACTORS. * indicates values consid-

ered significant at P 5 0.05, ** at P 5 0.01, NS 5 not significant after sequential Bonferroni correction.

FactorMean 6 SE

(msec) dfF

(ANOVA) H (K-W) P

Mandibular depressionMandibular elevationMandibular totalLabial onsetLabial protrusionLabial retractionLabial totalMaximum gapeFood enters mouth

26 6 166 6 492 6 48 6 1

27 6 265 6 492 6 432 6 236 6 2

444444444

3.958.88

4.282.48

11.5623.8624.7424.77

26.33

0.021*,0.001**,0.001**,0.001**

0.008*,0.001**,0.001**

0.005*0.056 NS

of the buccopharyngeal cavity during suctionfeeding.

In addition, up to 42 sequential bites fromindividual feeding events were analyzed asabove from two sharks that ate an unusual num-ber of food items (Nurse 4 5 80 cm TL, n 5 33bites; Nurse 5 5 79 cm TL, n 5 31 bites) tomeasure possible changes in duration of labialcartilage protrusion and retraction, and man-dibular depression and elevation as the sharksbecame satiated. These variables were consis-tently obtainable from the video sequences fornearly all of the bites in a feeding sequence.Video images that were not suitable for analysisresulted in the difference between the totalnumber of feeding bites (e.g., 42) and the num-ber of bites analyzed (e.g., 33). Satiation, whichoccurred in about 1–2 h, was determined by thesharks refusing to take additional food items.

Because the sharks were conditioned to feedwhen the flood lights were turned on, twosharks in particular (Nurse 4 and 5) often ap-proached the area where food was to be pre-sented and commenced suction feeding behav-ior in the absence of any food. This provided usthe opportunity to investigate stereotypy infeeding kinematics by comparing the durationof specific kinematic events (mandibular de-pression, mandibular elevation, mandibular to-tal, labial onset, labial protrusion, labial retrac-tion, labial total, maximum gape) of bites withfood present versus absent. The bites with andwithout food occurred throughout the entirefeeding trial and therefore were most likely notaffected by any satiation effects on duration (seeResults). In this manner, 12 bites with, and 12bites without food were compared separately foreach of the two sharks.

To construct a kinematic profile of cranialmovements during suction capture, sequentialvideo fields from a lateral view of a suction cap-

ture event were individually captured by theRedlake PCI-1000 software and analyzed withSigma Scan Pro (SPSS Science). The followingmeasurements were digitized every other field(0.008 sec) from the start of mandibular de-pression until the end of labial cartilage retrac-tion (totaling 48 fields): (1) labial cartilage pro-trusion distance 5 the distance from the pupilto anterior margin of the medial labial cartilage;(2) gape distance 5 distance from the anteriormargin of the mandible to anterior margin ofupper labia; (3) hyobranchial depression dis-tance 5 distance between the dorsal and ventralbody surface one third the distance from theeye to the first gill slit; (4) mandibular depres-sion and (5) cranial elevation angles 5 anglesformed from the ventral margin of the first gillslit, through the pupil, to the anterior tip of themandible and snout, respectively.

Statistical analysis.—Normality and equality ofvariances of each of the nine kinematic data setswere tested with the Kolmogorov-Smirnov testand the Levene Median test, respectively. Du-ration of each of the nine kinematic variables(e.g., mandibular depression) was then sepa-rately analyzed with a one-way ANOVA for dif-ferences among sharks (Table 1). Head eleva-tion occurred so infrequently as to preclude itsanalysis. If a difference was detected by ANOVA,a Tukey test was used to test all pair wise com-parisons. When normality or equality of varianc-es could not be achieved by transformation, aKruskal-Wallis one-way ANOVA on ranks wasused. If this indicated significant differencesamong sharks, a Dunn’s multiple comparisonsprocedure tested all pair wise comparisons (P ,0.05). Because many of these variables are cor-related, a sequential Bonferroni correction wasapplied to the ANOVA or Kruskal-Wallis analy-ses (Rice, 1989).

28 COPEIA, 2002, NO. 1

To test whether duration of capture kinemat-ics changes with satiation, duration of mandib-ular depression, mandibular elevation, labialprotrusion, and labial retraction were linearlyregressed against bite number independentlyfor two sharks that took numerous bites perfeeding (males, TL 79 and 80 cm). In the casewhere the shark took a piece of food but a clearvideo image of those particular bites was lack-ing, the regression analysis, which requires abalanced design, was performed without thosedata points (represented by gaps on the fig-ures). ANOVA was performed on the regres-sions after normality and equality of varianceswas demonstrated. The duration of mandibularelevation and labial retraction for Nurse 4 failednormality but passed equality of variances re-gardless of transformation. Therefore, in addi-tion to ANOVA, a Model II geometric mean(GM) regression on these data was calculated(Teissier, 1948), confirming the results of thelinear regression.

To test whether any change in duration ofmandibular depression or labial cartilage pro-trusion was a result of the shark changing theextent of mouth opening with satiation, the dis-tance from the eye to the anterior edge of fullyprotracted medial labial cartilage was regressedagainst bite number. As labial cartilage protrac-tion is mechanically linked to mandibular de-pression (Motta and Wilga, 1999), and the man-dible tip was not as clearly defined in the video,labial cartilage protraction was also used as aproxy for the extent of mandibular depression.Select video images were captured from videoand measured with Sigma Scan Pro (SPSS Sci-ence).

To compare the durations of the eight kine-matic variables with and without food, paired t-tests were conducted for each kinematic vari-able independently for each of the two sharkswith a sequential Bonferroni correction (Rice,1989). Statistical tests were performed with Sig-ma Stat ( Jandell Scientific Inc., vers. 2.0), withthe exception of the Bonferroni correction andGM regression, which were calculated indepen-dently.

RESULTS

Food capture.—Sharks were quickly conditionedto feed when the filming lights were turned on.In trials with the feeding platform, there weretwo patterns of food capture. In most feedingtrials, the shark would swim to the platform,stop before the food, raise itself on its pectoralfins, and suck the food item into its mouth. Inother trials, the shark used an apparent che-

mosensory-mediated search to find the food. Itwould swim slowly by the food, swinging its headfrom side to side, turn sharply, and suck it in.All food capture was by inertial suction in thatthe shark rapidly opened its mouth, and the ve-locity of the food toward the mouth greatly ex-ceeded the velocity of the shark approachingthe food, if the shark was moving at all. Whenconditioned to feed through a hole in the feed-ing platform, the shark would immediately swimonto the platform when the lights were turnedon, cease swimming, and in many cases beginnumerous suction attempts through the holeeven before food was presented. Some suctionattempts that appeared more vigorous were ac-companied by loud and audible poppingsounds perhaps indicating cavitation. The smallpieces of food offered in this experiment wereusually captured and swallowed in one suctionevent without being grasped by the teeth. Insome cases, the food was captured by inertialsuction, grasped by the teeth, and transportedfrom the buccal cavity into the esophagus by asecond suction event. Some pieces of food wereheld in the teeth or mouth after inertial suctioncapture, and repeatedly blown out of the mouthand sucked in again, reorienting the food orreducing it in size.

Food capture began with an expansive phase,which involved mandibular depression, labialcartilage protrusion, and occasionally cranial el-evation. The sequence began with mandibulardepression averaging 26 msec in duration (Figs.2–4, Table 1). Cranial elevation (29 msec) anddepression (29 msec) were uncommon andgenerally very slight, occurring in only 15% ofthe bites. In some cases, the head was actuallylowered prior to the bite as it was positionedtoward the food. The labial cartilages began toprotrude anteriorly eight msec after the begin-ning of mandibular depression. Maximum la-bial cartilage protrusion (27 msec) occurredslightly after maximum gape (32 msec), shortlyafter which the food was usually sucked into thebuccal cavity (36 msec) and no longer visible.The external gill slits closed during the expan-sive phase. The compressive phase began withmandibular elevation that took an average of 66msec. Palatoquadrate protrusion, obscured bylabial cartilage protrusion in perfectly lateralbites, began during the compressive phase andreached a maximum at approximately the sametime mandibular elevation ceased (Fig. 2). La-bial cartilage retraction (65 msec) then fol-lowed the beginning of mandibular elevation.As the compressive phase proceeded, the hyoidcomplex (basihyal and distal ceratohyal) was ob-served to bulge ventrally, then become ob-

29MOTTA ET AL.—NURSE SHARK FEEDING

Fig. 2. Food capture sequence of a 85 cm TL female Ginglymostoma cirratum suction feeding on food item. StartMD 5 start of mandibular depression as the food is visible anterior to the mouth; the caudal fin is visible behindthe food; max MD 5 maximum mandibular depression occurs as the food is entering the mouth; max lab prot 5maximum labial protrusion; max gape—FNLV 5 maximum gape and when the food is sucked into the mouth andno longer visible; end ME 5 the end of mandibular elevation; the protruded upper jaw (arrow), which appearswhite, is clearly visible in the mouth; max gill 5 maximum opening of the gill slits as the water rushes posteriorlyout of the branchial cavity. Numbers indicate msec elapsed from the beginning of mandibular depression.

30 COPEIA, 2002, NO. 1

Fig. 3. Composite diagram of means of kinematicevents for food capture (n 5 5 sharks, 53 bites) inGinglymostoma cirratum. Events (black bars) indicatestart to peak activity followed by peak to end of activity(grey bars), with 1 SE bars. Error bars on the left ofhead movement, labial cartilage movement, maxi-mum gape, and time until food enters the mouth in-dicate the SE of the onset time from the beginningof mandibular depression until initiation of thatevent. The dashed vertical line on left indicates theend of mandibular depression, and the dashed lineon the right indicates the cessation of the bite markedby retraction of the labial cartilages. Cranial elevationonly occurred in 15% of recorded bites and is notincluded in the statistical analysis.

scured by a general swelling of the entire pha-ryngeal cavity as the water preceded posteriorly,with the water eventually exiting through theopen external gill slits. The compressive phaseterminated with labial cartilage retraction atwhich time the protruded palatoquadrate wasstill clearly visible. The entire capture sequencefrom the beginning of mandibular depressionuntil retraction of the labial cartilages had anaverage duration of 100 msec. Palatoquadrateretraction continued into the recovery phase.

Individual variability.—There was considerablecombined within and among individual vari-ability in the timing of most kinematic eventsfor the five sharks (Fig. 5). Eight of the kine-matic variables involving mandibular depressionand elevation, labial cartilage protrusion and re-traction, and time to maximum gape exhibitedinterindividual variability based on the ANOVAresults, but there was no difference in the timeit took the food to enter the mouth (Table 1).

Effects of satiation on bite duration.—For one ofthe two sharks with multiple recorded bites(Nurse 4, male, TL 80 cm) the duration of la-bial cartilage protrusion (slope 520.37, F 55.11, P 5 0.031) and mandibular depression(slope 5 20.45, F 5 14.59, P 5 0.001) de-creased as the shark consumed more than 40pieces of food to satiation, that is, the shark

opened its mouth quicker with successive cap-tures (Fig. 6). However, the distance that thelabial cartilage extended anteriorly did notchange with increasing number of bites (F 51.65, P 5 0.208; therefore the quicker mouthopening was not reflective of the mouth simplyopening less. This shark did not change thetime to close the mouth as it became satiated,because there was no relationship between bitenumber and duration of mandibular elevation(F 5 1.61, P 5 0.29) or labial cartilage retrac-tion (F 5 0.43, P 5 0.515). For the other shark(Nurse 5), there was no change in the durationof mandibular depression and elevation or la-bial cartilage protrusion and retraction with in-creasing satiation (mandibular depression F 51.45, P 5 0.240; mandibular elevation F 5 1.03,P 5 0.321; labial cartilage protrusion F 5 2.86,P 5 0.101; labial cartilage retraction F 5 3.58,P 5 0.069).

Effects of food on bite durations.—For one shark(Nurse 4), there was essentially no difference inthe duration of the various kinematic events inthe presence or absence of food, except for thetime to elevate the mandible and the relatedtotal time to depress and elevate the mandible(Table 2). However, the other shark (Nurse 5)was consistently slower in the absence of food,with the exception of the time to depress themandible and the time for onset of labial car-tilage protrusion relative to the initiation ofmandibular depression (Table 2).

DISCUSSION

Conservation of kinematic sequence.—The inertialsuction feeding G. cirratum captures food withan abbreviated kinematic feeding sequencecompared to that of squaliform, carcharhini-form and lamniform sharks (Tricas and Mc-Cosker, 1984; Motta et al., 1991, 1997; Wilgaand Motta, 1998a). The feeding sequence of G.cirratum is composed of expansive, compressiveand recovery phases, with no evidence of an ini-tial mouth closing preparatory phase as hasbeen noted for some aquatic feeding verte-brates (Liem, 1978; Lauder, 1985; Lauder andPrendergast, 1992), although lacking buccalpressure profiles it is not possible to ascertaindefinitively the lack of a preparatory phase. Thesequence is further abbreviated by the lack ofhead lifting during the expansive phase.

The expansive phase of G. cirratum is char-acterized by mandibular depression and labialcartilage protrusion. As in the inertial suctionfeeding H. francisci and T. semifasciata (Ferry-Graham, 1998; Pretlow-Edmonds, 1999; Table

31MOTTA ET AL.—NURSE SHARK FEEDING

Fig. 5. Box plot of kinematic variables for all Gin-glymostoma cirratum combined (n 5 5 sharks and 53bites total per variable) showing median, 10th, 25th,75th, 90th percentiles, and outliers that exceed the90th percentile.

Fig. 4. Kinematic profile of an 85 cm TL femaleGinglymostoma cirratum suction feeding on a food item.Although this bite was longer in duration than theaverage bite, the shark was lateral to the camera fa-cilitating digitization. Zero msec indicates the begin-

←

ning of mandibular depression. Rapid mandibular de-pression is followed by labial cartilage protrusion andpeak gape. The mandibular depression trace ceasesafter 200 msec as the tip of the mandible is obscuredbecause of the shark turning slightly away from thecamera. The labial cartilages which were slightly pro-truded at the beginning of the capture consequentlyreturn to a more retracted position. Although cranialelevation only occurs in approximately 15% of suctioncaptures (Fig. 3), it is evident in this feeding event asthe head is elevated and slowly returns to its originalposition. Peak hyobranchial depression occurs towardthe end of the capture as the water rushes posteriorlythrough the mouth and pharynx and out the gill slits.The longer duration of this capture is primarily a re-sult of lengthening of the compressive and recoveryphases as mandibular depression occurs within ap-proximately 40 msec. The smaller scale oscillations inthe traces are a result of measurement error.

3), the expansive phase of G. cirratum lacks cra-nial elevation during the majority of bites, andwhen it does occur it is of small magnitude. Us-ing compensatory suction and ram, lamniformand carcharhiniform sharks often consumelarge prey items with their ventrally locatedmouth. In these sharks cranial elevation is cou-pled with mandibular depression to open themouth as much as possible and direct the gapemore anteriorly toward the prey. However, G.cirratum and H. francisci primarily capture rela-tively small fishes and benthic invertebrates(Bigelow and Schroeder, 1948; Compagno,1984; Castro, 2000) with a mouth that is almostterminal when maximally open (Edmonds etal., 2001). Lifting of the cranium to orient theopen mouth is not necessary. Conversely, whenfeeding on small benthic food items, T. semifas-ciata swims over them protruding its tubular

32 COPEIA, 2002, NO. 1

Fig. 6. Linear regressions of labial cartilage pro-trusion, and mandibular depression as a function ofnumber of consecutive bites for (A) nurse 4, a 80 cmTL male fed up to 42 pieces of food during one ex-periment, and (B) nurse 5, a 79 cm TL male Gingly-mostoma cirratum fed up to 39 pieces of food duringone experiment. For the first shark the time for themandible to depress (P 5 0.001) and the time for thelabial cartilage to protrude (P 5 0.031) both de-creased with satiation. Because the distance these el-ements moved did not change, the velocity of jawopening increased with satiation.

mouth anteroventrally to capture them, result-ing in little if any cranial elevation (Ferry-Gra-ham, 1998).

During the compressive phase the mandibleis elevated, the palatoquadrate protruded, thehyoid depressed and the labial cartilages retract-ed. Not only is mandibular depression in G. cir-ratum the fastest recorded for any shark, but incontrast to nearly all sharks examined so far, theexpansive phase in G. cirratum is considerablyfaster than the compressive phase (Table 1).Therefore, similar to some other bony fishes,aquatic amphibians, and some aquatically feed-ing amniotes, G. cirratum has a fast openingphase during the expansive phase (Lauder,

1985; Lauder and Reilly, 1994; Summers et al.,1998).

The recovery phase in carcharhiniform andsqualiform sharks includes retraction of the pal-atoquadrate, hyoid, and branchial apparatusinto their resting position (Motta et al., 1997;Wilga and Motta, 1998a). In G. cirratum the pal-atoquadrate is retracted; however, the generalbulging of the buccal and branchial cavities re-sulting from water inflow masks movements ofthe hyoid and branchial apparatus in the videoimages. Presumably the hyoid is retracted to itsresting position. This awaits confirmation byour electromyographic experiments.

Water flow during inertial suction feeding inG. cirratum is apparently unidirectional, in themouth and out of the gill slits, as evidenced bythe posteriorly traveling bulge in the head asthe mouth closes and the gill slits open duringthe compressive phase, and the position of thebuccal valves at the anterior end of the mouth(Motta and Wilga, 1995). This contrasts with wa-ter flow in juvenile swell sharks Cephaloscylliumventriosum, during ram-dominated feeding inwhich water taken in the mouth is later forcedout the mouth rather than continuing posteri-orly through the gill slits (Ferry-Graham, 1997).The reverse flow of water out the mouth of G.cirratum is hindered by two passive connectivetissue oral valves at the anterior end of themouth. These valves, which are lacking in thecarcharhinid N. brevirostris, are extensions of themandibulohyoid connective tissue sheath (Mot-ta and Wilga, 1995, 1999).

Most sharks capture their prey by inertial suc-tion, compensatory suction, ram or a combina-tion of these, or by biting a piece off a preyitem. After capture there is often one or moremanipulation bites in which the prey is reducedin size by cutting with the teeth, followed bytransport of the bolus of food from the mouthinto the esophagus (Motta and Wilga, 2001). Inall sharks examined to date, transport is alwaysby suction (Ferry-Graham, 1997; Motta et al.,1997; Wilga and Motta, 1998a). Under the ex-perimental conditions reported here, andthrough observation of juvenile G. cirratumfeeding on small pieces of food in the field(PJM, pers. obs.), the food is usually capturedand transported by one initial inertial suctionevent. However, in some cases, pieces of foodare captured by inertial suction and held in theteeth, whereupon larger pieces may be repeat-edly blown out of the mouth and sucked in, dis-membering them, and then subsequently trans-ported by suction from the mouth into theesophagus.

Lauder (1980b, 1983) proposed a model of

33MOTTA ET AL.—NURSE SHARK FEEDING

TABLE 2. KINEMATIC VARIABLES WITH AND WITHOUT FOOD FOR TWO Ginglymostoma cirratum AND RESULTS OF THEPAIRED t TESTS. * indicates values considered significant at P 5 0.05, ** at P 5 0.01, NS 5 not significant after

sequential Bonferroni correction.

Nurse 4 Factor

Duration withfood 6 SE

(msec)

Duration w/ofood 6 SE

(msec) df t value P

Mandibular depressionMandibular elevationMandibular totalLabial onsetLabial protrusionLabial retractionLabial totalMaximum gape

29 6 383 6 6

113 6 96 6 1

30 6 382 6 8

113 6 935 6 3

40 6 5128 6 5167 6 910 6 238 6 3

115 6 4154 6 648 6 5

1111111111111111

21.41824.88523.65921.29521.74322.91022.97921.951

0.184 NS,0.001**

0.004*0.222 NS0.109 NS0.014 NS0.013 NS0.077 NS

Nurse 5 Factor

Duration withfood 6 SE

(msec)

Duration w/ofood 6 SE

(msec) df t value P

Mandibular depressionMandibular elevationMandibular totalLabial onsetLabial protrusion

22 6 277 6 398 6 32 6 0

25 6 2

30 6 3102 6 4132 6 5

4 6 134 6 3

1111111111

22.24625.70225.87520.71324.158

0.046 NS,0.001**,0.001**

0.491 NS0.002**

Labial retractionLabial totalMaximum gape

78 6 2103 6 225 6 2

109 6 5143 6 639 6 3

111111

26.42327.40723.484

,0.001**,0.001**

0.005*

inertial suction feeding in bony fishes that ap-proach their food slowly or remain stationarybefore food capture. The gill arches effectivelyseparate the buccal and parabranchial/opercu-lar cavities such that a large pressure differentialis set up between the two cavities during whichthe food is sucked into the mouth. Abductionof the operculum has a negligible role in gen-erating subambient mouth cavity pressures. Af-ter peak subambient buccal pressure is reached,the abducting opercula result in an anteropos-terior flow of water from the buccal cavitythrough the gills into the opercular cavity, afterwhich it exits through the opercular slit. Thismechanism is contingent on a rigid operculumthat is forcibly abducted. This mechanism can-not function during inertial suction feeding insharks as they lack a rigid operculum. Subam-bient pressures generated external to the gillfilaments in the parabranchial chamber ofsharks during respiration are apparently a resultof elastic recoil of the visceral skeleton and notactive expansion of this cavity (Hughes and Bal-lintijn, 1965; Ferry-Graham, 1999). Therefore,the suction mechanism of G. cirratum is bestmodeled by an expanding single buccopharyn-geal/orobranchial cavity resulting from depres-sion of the mandible, depression of the floor ofthe mouth (hyoid apparatus) and depression ofthe branchial region. This is followed by reduc-tion of the cavities resulting from mandibular

elevation and subsequent hyoid and branchialelevation that results in the anteroposterior flowof water from the mouth out of the parabran-chial chambers.

Variability.—Many bony fishes exhibit interindi-vidual variability of motor and kinematic pat-terns within food types and can modulate theirfeeding mechanisms in response to differentprey/food types and positions (Liem, 1978; San-derson, 1988; Nemeth, 1997). To date, the re-sults on modulatory abilities of elasmobranchfishes are conflicting, but interindividual vari-ability in food capture kinematics appears to bewidespread in sharks (reviewed in Motta andWilga, 2001). There was interindividual vari-ability in all of the kinematic variables in G. cir-ratum with the exception of the time requiredfor the food to enter the mouth, which aver-aged 36 msec. Despite the variability, the kine-matic sequence during food capture was veryconsistent and stereotyped. Even in the absenceof food, one shark displayed no difference inthe majority of kinematic durations as com-pared to bites with food, and both sharks didnot differ in the time to depress the mandiblewith or without food present. Rapid mandibulardepression is an important parameter in thegeneration of inertial suction forces (Lauder,1980a; Muller et al., 1982; Liem, 1993). A studyof feeding kinematics in G. cirratum ranging in

34 COPEIA, 2002, NO. 1

TA

BL

E3.

ME

AN

CA

PTU

RE

BIT

ED

UR

AT

ION

S(M

SEC

)FO

RSE

LE

CT

KIN

EM

AT

ICV

AR

IAB

LE

SFO

RA

VA

RIE

TY

OF

SHA

RK

S.Pr

edom

inan

tfo

odca

ptur

ebe

havi

or(r

amvs

iner

tial

suct

ion)

indi

cate

dbe

low

each

spec

ies.

Inco

mpl

ete

valu

esin

dica

teth

ose

data

are

not

avai

labl

e(N

/A).

G.c

irra

tum

(suc

tion)

H.f

ranc

isci

(suc

tion)

aT.

sem

ifasc

iata

(suc

tion-

ram

)bS.

acan

thia

s(s

uctio

n-ra

m)c

S.tib

uro

(ram

)dN

.bre

viro

stris

(ram

)eC.

pere

zi(r

am)f

C.ve

ntrio

sum

(ram

)gC.

carc

haria

s(r

am)h

Hea

del

evat

ion

Hea

dde

pres

sion

Man

dibu

lar

depr

essi

onM

andi

bula

rel

evat

ionj

Max

imum

prot

rusi

onl

Max

imum

gape

Tota

lbite

time

rare

rare 26 66 N/A 32 100

rare

rare 47

–64

66–8

421

–25

47–6

411

3–14

8

none

orsl

ight

none

orsl

ight

90–1

1060

–70

80–1

10

90–1

1015

0–18

0

89–1

6512

9–15

065

–111

58–1

3746

–69

100

266–

280

111

113

162 86 57 162

302

66 62 75 77 59 81 309

84 83 115

105 53 120

383

none

226–

334i

305–

354

N/A

k

none

orsl

ight

236–

347

306–

348

367–

419

171

405

140

220 83 167

405

aPr

etlo

w-E

dmon

ds(1

999)

for

thre

efo

odpr

esen

tatio

ns.

bFe

rry-

Gra

ham

(199

8)fo

rtw

ofo

odty

pes

and

two

food

size

s.c

Wilg

a(1

997)

for

thre

efo

odsi

zes.

dW

ilga

(199

7).

eM

otta

etal

.(19

97).

fM

otta

and

Wilg

a(2

001)

.g

Ferr

y-G

raha

m(1

997)

for

two

food

size

s.h

Tric

asan

dM

cCos

ker

(198

4),T

rica

s(1

985)

.iT

ime

tom

axim

umne

gativ

ehe

adan

gle.

jFr

ompe

akde

pres

sion

.k

Dur

ing

capt

ure

food

was

held

with

the

teet

hre

sulti

ngin

noja

wcl

osur

e.lD

urat

ion

from

star

tof

uppe

rja

wpr

otru

sion

until

max

imum

prot

rusi

on.

35MOTTA ET AL.—NURSE SHARK FEEDING

size from 33–268 cm TL similarly found that theprey capture sequence did not change ontoge-netically, the angular and linear kinematic ex-cursions of the jaw elements were isometric, andthe use of inertial suction feeding as judged bythe Ram Suction Index (Norton and Brainerd,1993) did not change with size (Robinson andMotta, 2002).

As G. cirratum became satiated, the time toopen the mouth was very consistent for one ofthe sharks but decreased for the second. Thefact that the labial cartilages extended the samedistance for the latter shark indicated that thiswas not simply an artifact of the shark openingthe mouth less, but rather the velocity of man-dibular depression increased. This contrastswith Lepomis sp. (Perciformes: Centrarchidae)that decrease buccal pressure magnitude withsatiation (Lauder, 1980b), and largemouth bassMicropterus salmoides (Centrarchidae) that de-crease the speed of mouth opening and alsoopen the mouth less as they become satiated(Sass, 1999). The ecological implications of theinterindividual variability in kinematic eventsand the effects of satiation on bite speed areprobably minimal in G. cirratum as the suctionevents are so rapid and powerful that the foodenters the mouth in the same time regardlessof the variability, and the time to mandibulardepression in one of the satiated G. cirratumonly changed by approximately 20 msec. How-ever, how such differences in kinematic eventsmay affect capture success of whole prey remainto be tested.

The conservative feeding kinematics underthese experimental conditions in which the sizeand type of food are held constant indicate thatinertial suction food capture in G. cirratum is astereotyped modal action pattern (sensu Barlow1968, 1977). This implies that the bites are verystereotyped, involve central nervous system in-tegration (not a simple reflex), are indepen-dent from feedback once initiated, involvegreater variability in the taxic component thanin the kinetic component, and the behavior iswidely distributed among individuals of a pop-ulation (i.e., heritable). Other orectolobiformsharks (O. maculatus and Hemiscyllium ocellatum)are also apparently stereotyped in their foodcapture kinematics (Wu, 1994). Heterodontusfrancisci (Heterodontiformes) showed stereo-typed suction feeding when presented food ofdiffering accessibility (Edmonds et al., 2001).Similarly, C. ventriosum and T. semifasciata (Car-charhiniformes) do not appear to modulatecapture kinematics when presented with food ofdifferent sizes (Ferry-Graham, 1997, 1998).Even though S. californica vary their attack ap-

proach based on position of the food in the wa-ter column, a stereotyped kinematic sequencewas used for capture. The bites of S. californicaare believed to represent a modal action pattern(Fouts and Nelson, 1999), as Tricas (1985) pro-posed for C. carcharias feeding on surface food.

In contrast, ram feeding carcharhinid sharksincluding the Caribbean reef shark Carcharhinusperezi, blacknose Carcharhinus acronotus, blacktipCarcharhinus limbatus, and N. brevirostris showconsiderable variation in the timing and extentof upper jaw protrusion during food capture(Frazzetta and Prange, 1987; Motta et al., 1997;Motta and Wilga, 2001). Squalus acanthias(Squaliformes) is capable of using a wide varietyof food capture strategies on a single food type,and there is extensive modulation between cap-ture and transport events (Wilga and Motta,1998a). This ability is properly referred to asvariability rather than modulation though, be-cause there is a continuous range of behaviorsin the face of a single stimulus (Ferry-Graham,1997). Despite the apparent lack of modulatoryability in some sharks, individual variability iscommon during feeding in all sharks examinedso far (Motta et al., 1997; Ferry-Graham, 1998;Wilga and Motta, 1998a).

These preliminary findings must be viewedwith caution as most of these studies were con-fined to laboratory conditions and a diverserange of food types, sizes, or prey mobilitieswere not tested. However, the emerging patternsuggests a continuum from more stereotypedand faster food capture bites in morphologicallyspecialized, and benthic inertial suction feedingsharks such as G. cirratum, to less stereotyped,and generally slower bites in larger ram feedingsharks (Table 3).

Specialization for inertial suction prey capture.—Suc-tion feeding is the ancestral dominant mode ofprey capture in bony fishes (Lauder, 1985).Bony fishes that are suited for inertial suctionfeeding usually have a small, laterally enclosedmouth, protrusible upper jaw, reduced denti-tion, strongly developed abductor muscles, andbuccal expansion is large and rapid (Lauder,1980a; Muller et al., 1982; Liem, 1993). The par-cel of water sucked into the mouth is usuallysmall and particles move toward the center lineof the mouth from each side of the head, al-though most of the water entering the mouthpasses through the area anterior to the gape(Lauder and Clark, 1984).

Specializations for inertial suction prey cap-ture in G. cirratum and inertial suction feedingbony fishes are functionally convergent. Theopen mouth of this shark is small (less than one-

36 COPEIA, 2002, NO. 1

third head length), anteriorly directed andclose to the tip of the snout. The functionallyterminal mouth is a derived character of orec-toloboids (Bigelow and Schroeder, 1948; Com-pagno, 1984, 1988). The open mouth is approx-imately round when open and laterally occlud-ed by three prominent labial cartilages and thepalatoquadratomandibular connective tissuesheath. A medially directed cartilaginous pro-cess of the medial cartilage abuts the upper jaw,holding the labial cartilages away from the jawand preventing their collapse during the pow-erful sub-ambient suction forces, which canreach 2760 mm Hg. Two buccal valves inhibitback-flow of water out of the mouth (Tanaka,1973; Motta and Wilga, 1999). Labial cartilagesthat laterally occlude the gape and form aroughly circular mouth are also found in theinertial suction feeding H. francisci (Pretlow-Ed-monds, 1999), T. semifasciata (Ferry-Graham,1998), and S. acanthias (Wilga and Motta,1998a). Mandibular depression in G. cirratum isextremely rapid (26 msec) and the fastest re-corded for an elasmobranch (Table 3). The de-pressor muscles of the hyoid and branchialarches, most notably the coracohyoideus andthe coracobranchiales, are apparently hypertro-phied compared to carcharhinid sharks (Moss1965, 1977). Ginglymostoma cirratum comes closeto functionally replicating the feeding mecha-nism of an inertial suction feeding bony fish(Lauder, 1980b; Liem, 1980b), and it does itwith a similar suite of modifications to the preycapture apparatus.

Fossil evidence and observation of morpho-logically similar but extant sharks indicate theancestral prey capture behavior in elasmo-branchs might have involved grasping the preywhole or tearing pieces from it (Schaeffer, 1967;Moy-Thomas and Miles, 1971; Maisey, 1980).Specialization for inertial suction prey capturehas apparently arisen repeatedly in sharks andrays, often nested within clades of ram feeders(Fig. 1). Therefore, specialization for inertialsuction feeding including rapid jaw opening,formation of a round, somewhat terminalmouth, prominent labial cartilages, a dentitionreduced in size, and generation of large sub-ambient suction pressures has apparentlyevolved independently in conjunction with abenthic lifestyle in several elasmobranch line-ages. These sharks and rays feed on both elusiveand non elusive prey that lives in or on the sub-strate, are attached to it, or are benthic-associ-ated (Compagno, 1984). Rhincodon typus is a no-table exception. This shark has a relatively largeterminal mouth and slow jaw kinematics, andutilizes a pulsatile suction and filtering mecha-

nism to feed on a wide variety of planktonic andnektonic prey that is generally aggregated inspace and much smaller relative to its body size(Taylor et al., 1983; Clark and Nelson, 1997;Colman, 1997).

In summary, the inertial suction feeding G.cirratum captures its food with a conservativebut abbreviated kinematic sequence of cranialmovements. Water flow during inertial suctionfeeding is unidirectional, and the suction mech-anism is driven by the expanding buccopharyn-geal cavity. Despite interindividual variability inthe feeding kinematics, inertial suction foodcapture in this shark is apparently a stereotyped,modal action pattern. Specializations for iner-tial suction feeding include a small, terminaland laterally occluded mouth, small teeth, mod-ified labial cartilages, oral valves to preventback-flow of water out of the mouth, hypertro-phied abductor muscles of the hyoid and bran-chial arches, and extremely fast buccal expan-sion. These functional and morphological spe-cializations make G. cirratum the cartilaginousfish equivalent of the bony fish inertial suctionspecialist. These specializations are shared tovarious degrees with other shark species, includ-ing members of several other families. This spe-cialization for inertial suction feeding in sharksand bony fishes represents an important exam-ple of functional convergence.

ACKNOWLEDGMENTS

We gratefully acknowledge the contributionsof time and material made by many persons andinstitutions. S. Barker, R. Carr, N. Edwards, S.Gold, A. Griffith, E. Minor, T. Pietrzak, L. Riley,C. Wilga and C. Wood provided technical assis-tance. E. Sander, J. Bonnel, C. Luer, Keys Ma-rine Laboratory and Mote Marine Laboratorydonated specimens. Mote Marine Laboratoryand the University of South Florida provided fa-cilities and expertise. Animals were treated ac-cording to the University of South Florida andMote Marine Laboratory Institutional AnimalCare and Use Committee guidelines, Certificate510. The project was supported by grants fromthe National Science Foundation to PJM andREH (DEB 9117371 and IBN 9807863).

LITERATURE CITED

BARLOW, G. W. 1968. Ethological units of behavior. p.217–232. In: The central nervous system and fishbehavior. D. Ingle (ed.). Univ. of Chicago Press,Chicago.

———. 1977. Modal action patterns. p. 98–134. In:How animals communicate. T. A. Sebeok (ed.). In-diana Univ. Press, Bloomington.

37MOTTA ET AL.—NURSE SHARK FEEDING

BIGELOW, H. B., AND W. C. SCHROEDER. 1948. Fishesof the western North Atlantic. Lancelets, cyclo-stomes, sharks. Mem. Sears Found. Mar. Res.1(part1):1–576.

CARRIER, J. C., AND C. A. LUER. 1990. Growth rates inthe nurse shark, Ginglymostoma cirratum. Copeia1990:686–692.

CASTRO, J. I. 2000. The biology of the nurse shark,Ginglymostoma cirratum, off the Florida east coastand the Bahama Islands. Environ. Biol. Fish. 58:1–22.

CHU, C. T. 1989. Functional design and prey capturedynamics in an ecologically generalized surfperch(Embiotocidae). J. Zool. 217:417–440.

CLARK, E., AND D. R. NELSON. 1997. Young whalesharks, Rhincodon typus, feeding on a copepodbloom near La Paz, Mexico. Environ. Biol. Fish. 50:63–73.

COLMAN, J. G. 1997. A review of the biology and ecol-ogy of the whale shark. J. Fish Biol. 51:1219–1234.

COMPAGNO, L. J. V. 1984. Sharks of the world. An an-notated and illustrated catalogue of shark speciesknown to date. FAO Species Catalogue. Vol. 4. Part1. Hexanchiformes to Lamniformes. U.N. Dev.Prog., FAO, Rome.

———. 1988. Sharks of the order Carcharhiniformes.Princeton Univ. Press, Princeton, NJ.

DIAMOND, J. M. 1985. Filter-feeding on a grand scale.Nature 316:679–680.

EDMONDS, M. A., P. J. MOTTA, AND R. E. HUETER. 2001.Prey capture kinematics of the suction feedinghorn shark, Heterodontus francisci. Environ. Biol.Fish. 62:415–427.

FERRY-GRAHAM, L. A. 1997. Feeding kinematics of ju-venile swellsharks, Cephaloscyllium ventriosum. J. Exp.Biol. 200:1255–1269.

———. 1998. Effects of prey size and mobility onprey-capture kinematics in leopard sharks, Triakissemifasciata. Ibid. 201:2433–2444.

———. 1999. Mechanics of ventilation in swellsharks,Cephaloscyllium ventriosum (Scyliorhinidae). Ibid.202:1501–1510.

FOUTS, W. R., AND D. R. NELSON. 1999. Prey captureby the Pacific angel shark, Squatina californica: vi-sually mediated strikes and ambush-site character-istics. Copeia 1999:304–312.

FRAZZETTA, T. H., AND C. D. PRANGE. 1987. Movementsof cephalic components during feeding in some re-quiem sharks (Carcharhiniformes: Carcharhini-dae). Ibid. 1987:979–993.

HUGHES, G. M., AND C. M. BALLINTIJN. 1965. The mus-cular basis of the respiratory pumps in the dogfish(Scyliorhinus canicula). J. Exp. Biol. 43:363–383.

LAUDER, G. V. 1980a. Evolution of the feeding mech-anism in primitive actinopterygian fishes: a func-tional anatomical analysis of Polypterus, Lepisosteus,and Amia. J. Morphol. 163:283–317.

———. 1980b. The suction feeding mechanism insunfishes (Lepomis): an experimental analysis. J.Exp. Biol. 88:49–72.

———. 1983. Prey capture hydrodynamics in fishes:experimental tests of two models. Ibid. 104:1–13.

———. 1985. Aquatic feeding in lower vertebrates, p.210–229. In: Functional vertebrate morphology. M.Hildebrand, D. M. Bramble, K. F. Liem, and D. B.Wake (eds.). Belknap Press, Cambridge.

———. 1990. Functional Morphology and systemat-ics: Studying functional patterns in an historicalcontext. Ann. Rev. Ecol. Sys. 21:317–340.

———, AND B. D. CLARK. 1984. Water flow patternsduring prey capture by teleost fishes. J. Exp. Biol.113:143–150.

———, AND T. PRENDERGAST. 1992. Kinematics ofaquatic prey capture in the snapping turtle Chelydraserpentina. Ibid. 164:55–78.

———, AND S. M. REILLY. 1994. Amphibian feedingbehavior: Comparative biomechanics and evolu-tion, p. 163–195. In: Biomechanics of feeding invertebrates. V. Bels, M. Chardon, and P. Vandewalle(eds.). Springer-Verlag, Berlin, Germany.

LEBOEUF, B. J., J. E. MCCOSKER, AND J. HEWITT. 1987.Crater wounds on northern elephant seals: thecookiecutter shark strikes again. U.S. Fish. Bull. 85:387–392.

LIEM, K. F. 1978. Modulatory multiplicity in the func-tional repertoire of the feeding mechanisms in cich-lid fishes. I. Piscivores. J. Morphol. 158:323–360.

———. 1980a. Adaptive significance of intra- and in-terspecific differences in the feeding repertoires ofcichlid fishes. Am. Zool. 20:295–314.

———. 1980b. Acquisition of energy by teleosts: adap-tive mechanisms and evolutionary patterns, p. 299–334. In: Environmental physiology of fishes. M.A. Ali(ed.). Plenum Publishing Corporation, New York.

———. 1993. Ecomorphology of the teleostean skull,p. 422–452. In: The skull. Functional and evolution-ary mechanisms. J. Hanken, and B. K. Hall (eds.).Vol. 3. Univ. of Chicago Press, Chicago.

MAISEY, J. G. 1980. An evaluation of jaw suspension insharks. Am. Mus. Novit. 2706:1–17.

MOSS, S. A. 1965. The feeding mechanisms of threesharks: Galeocerdo cuvieri (Peron & Le Sueur), Ne-gaprion brevirostris (Poey), and Ginglymostoma cirra-tum (Bonnaterre). Unpubl. Ph.D. diss., CornellUniv., Ithaca, NY.

———. 1977. Feeding mechanisms in sharks. Am.Zool. 17:355–364.

MOTTA, P. J., AND C. D. WILGA. 1995. Anatomy of thefeeding apparatus of the lemon shark, Negaprion bre-virostris. J. Morphol. 226:309–329.

———, AND ———. 1999. Anatomy of the feedingapparatus of the nurse shark, Ginglymostoma cirra-tum. Ibid. 241:33–60.

———, AND ———. 2001. Advances in the study offeeding behaviors, mechanisms, and mechanics ofsharks. Environ. Biol. Fish. 60:131–156.

———, R. E. HUETER, AND T. C. TRICAS. 1991. An elec-tromyographic analysis of the biting mechanism ofthe lemon shark, Negaprion brevirostris: functionaland evolutionary implications. J. Morphol. 210:55–69.

———, T. C. TRICAS, R. E. HUETER, AND A. P. SUM-MERS. 1997. Feeding mechanism and functionalmorphology of the jaws of the lemon shark Nega-prion brevirostris (Chondrichthyes, Carcharhinidae).J. Exp. Biol. 200:2765–2780.

MOY-THOMAS, J. A., AND R. S. MILES. 1971. Paleozoicfishes. Chapman and Hall, London.

MULLER, M., J. W. M. OSSE, AND J. H. G. VERHAGEN.1982. A quantitative hydrodynamic model of suc-tion feeding in fish. J. Theor. Biol. 95:49–79.

NEMETH, D. H. 1997. Modulation of attack behavior

38 COPEIA, 2002, NO. 1

and its effect on feeding performance in a trophicgeneralist fish, Hexagrammos decagrammus. J. Exp.Biol. 200:2155–2164.

NORTON, S. F., AND E. L. BRAINERD. 1993. Conver-gence in the feeding mechanics of ecomorpholog-ically similar species in the Centrarchidae and Cich-lidae. Ibid. 176:11–29.

PRETLOW-EDMONDS, M. A. 1999. Prey capture kinemat-ics in the horn shark, Heterodontus francisci. Unpubl.master’s thesis, Univ. of South Florida, Tampa.

RICE, W. R. 1989. Analyzing tables of statistical tests.Evolution 43:223–225.

ROBINSON, M. P., AND P. J. MOTTA. 2002. Patterns ofgrowth and the effects of scale on the feeding ki-nematics of the nurse shark, Ginglymostoma cirratum.J. Zool. 256.

RUSSO, R. A. 1975. Observations on the food habitsof leopard sharks (Triakis semifasciata) and brownsmooth-hounds (Mustelus henlei). Calif. Fish Game61:95–103.

SANDERSON, S. L. 1988. Variation in neuromuscularactivity during prey capture by trophic specialistsand generalists (Pisces: Labridae). Brain Behav.Evol. 32:257–268.

SASS, G. G. 1999. The effects of satiation on prey cap-ture kinematics in the largemouth bass, Micropterussalmoides. Unpubl. honors thesis, Univ. of SouthFlorida, Tampa.

SCHAEFFER, B. 1967. Comments on elasmobranch evo-lution. p. 3–35. In: Sharks, skates and rays. P. W.Gilbert, R. F. Mathewson, and D. P. Rall (eds.).Johns Hopkins Press, Baltimore, MD.

SHIRAI, S. 1996. Phylogenetic interrelationships ofneoselachians (Chondrichthyes: Euselachii), p. 9–34. In: Interrelationships of fishes. M. L. J. Stiassny,L. R. Parenti, and G. D. Johnson (eds.). AcademicPress, New York.

SIMS, D. W. 2000. Filter-feeding and cruising swim-ming speeds of basking sharks compared with op-timal models: they filter-feed slower than predictedfor their size. J. Exp. Mar. Biol. Ecol. 249:65–76.

STRONG JR., W. R. 1989. Behavioral ecology of hornsharks, Heterodontus francisci, at Santa Catalina Is-land, California, with emphasis on patterns of spaceutilization. Unpubl. master’s thesis, California StateUniv., Long Beach.

SUMMERS, A. P., K. F. DAROUIAN, A. M. RICHMOND, ANDE. L. BRAINERD. 1998. Kinematics of aquatic and ter-restrial prey capture in Terrapene carolina, with im-plications for the evolution of feeding in cryptodireturtles. J. Exp. Zool. 281:280–287.

TALENT, L. G. 1976. Food habits of the leopard shark,Triakis semifasciata, in Elkhorn Slough, MontereyBay, Calif. Fish Game. 62:286–298.

TANAKA, S. K. 1973. Suction feeding by the nurseshark. Copeia 1973:606–608.

TAYLOR, L. R., L. J. V. COMPAGNO, AND P. J. STRUH-SAKER. 1983. Megamouth—a new species, genus,and family of lamnoid shark (Megachasma pelagios,family Megachasmidae) from the Hawaiian Islands.Proc. Calif. Acad. Sci. 43:87–110.

TEISSIER, G. 1948. La relation d’allometrie: sa signifi-cation statistique et biologique. Biometrics 4:14–43.

TRICAS, T. C. 1985. Feeding ethology of the whiteshark, Carcharodon carcharias. Mem. sth. Calif. Acad.Sci. 9:81–91.

———, AND J. E. MCCOSKER. 1984. Predatory behaviorof the white shark (Carcharodon carcharias), withnotes on its biology. Proc. Calif. Acad. Sci. 43:221–238.

VAN DAMME, J., AND AERTS, P. 1997. Kinematics andfunctional morphology of aquatic feeding Austra-lian side necked turtles (Pleurodira: Chelodina). J.Morphol. 233:113–125.

WAINWRIGHT, P. C., AND R. G. TURINGAN. 1993. Cou-pled versus uncoupled functional systems: motorplasticity in the queen triggerfish Balistes vetula. J.Exp. Biol. 180:209–227.

WILGA, C. A. 1997. Evolution of feeding mechanismsin elasmobranchs: a functional morphological ap-proach. Unpubl. Ph.D. diss., Univ. of South Florida,Tampa.

WILGA, C. D., AND P. J. MOTTA. 1998a. Conservationand variation in the feeding mechanism of thespiny dogfish Squalus acanthias. J. Exp. Biol. 201:1345–1358.

———, AND ———. 1998b. Feeding mechanism ofthe Atlantic guitarfish Rhinobatos lentiginosus: mod-ulation of kinematic and motor activity. Ibid. 201:3167–3184.

———, AND ———. 2000. Durophagy in sharks: feed-ing mechanics of the hammerhead Sphyrna tiburo.Ibid. 203:2781–2796.

———, P. C. WAINWRIGHT, AND P. J. MOTTA. 2000. Evo-lution of jaw mechanics in vertebrates: insightsfrom Chondrichthyes. Biol. J. Linn. Soc. 71:165–185.

WU, E. H. 1994. A kinematic analysis of jaw protrusionin orectolobiform sharks: a new mechanism for jawprotrusion in elasmobranchs. J. Morphol. 222:175–190.

(PJM) DEPARTMENT OF BIOLOGY, UNIVERSITY OFSOUTH FLORIDA, 4202 EAST FOWLER AVENUE,TAMPA, FLORIDA 33620; (REH) CENTER FORSHARK RESEARCH, MOTE MARINE LABORATORY,1600 KEN THOMPSON PARKWAY, SARASOTA,FLORIDA 34236; (TCT) DEPARTMENT OF BIO-LOGICAL SCIENCES, FLORIDA INSTITUTE OFTECHNOLOGY, 150 WEST UNIVERSITY BOULE-VARD, MELBOURNE, FLORIDA 32901; AND (APS)MUSEUM OF VERTEBRATE ZOOLOGY, DEPART-MENT OF INTEGRATIVE BIOLOGY, UNIVERSITY OFCALIFORNIA, BERKELEY, CALIFORNIA 94720.PRESENT ADDRESSES: (TCT) DEPARTMENT OFZOOLOGY, UNIVERSITY OF HAWAII AT MANOA,2538 MCCARTHY HALL, HONOLULU, HAWAII96822; AND (APS) DEPARTMENT OF ECOLOGYAND EVOLUTIONARY BIOLOGY, UNIVERSITY OFCALIFORNIA, IRVINE, CALIFORNIA, 92697. E-mail: (PJM) motta@chuma1.cas.usf.edu. Sendreprint requests to PJM. Submitted: 23 May2000. Accepted: 11 June 2001. Section editor:J. D. McEachran.