Predatory Behavior of White Sharks

7/27/2019 Predatory Behavior of White Sharks

1/15

] AIm: Bio!. Ass. UEl. (2005),85,1121- 1135Printed in the United Kingdom

Predatory behaviour of white sharks (Carcharodon carcharias) at Seal Island, South Africa

R. Aidan M a r t i n * t ~ l , Neil Hammerschlagt+, Ralph S. CollierS and Chris Fallows*Fish Museum, Zoology Department, University of British Columbia, 6270 University Boulevard, Vancouver, BC ,V6T IZ4, Canada.

tReefQuest Centre for Shark Research, PO Box 485 61, 595 Burrard Street, Vancouver, BC ,V7X IA 3, Canada. lPew Institute for Ocean Science, Rosenstiel School of Marine and Atmospheric Sc ience, University of Miami, Miami , FL 33149, USA.

JShark Re search Committee, PO Box 3483, Van Nuys, CA 97407, USA. Apex Expeditions, 14 Thibault Walk , Marina Da Gama, 7945 Cape Town , South Africa.

e s p o n d author, e-mail: ram @elasmo-research.org

Between 1997 and 2003, there were 2088 natural predations by white sharks (Carcharodon carcharias) onCape fur seals (Arctocephalus pusillus pusillus) and 121 strikes on towed seal-shaped decoys were documentedfrom observation vessels at Seal Island, South Africa. White sha r.ksat Seal Island appear to selectivelytarget lone, incoming young of the year Cape fur seals at or near the surface. Most attacks lasted < I minand consisted of a single breach, with predatory success rate decreasing rapidly with increasing durationand number of subsequenfbreaches. A white shark predatory ethcigram, co mposed of four phases and 20behavioural units, is presented, including four va rieties of initial str ike and 11 subsei:J.uent behaviour unitsnot previously defined in the literature. Behaviour units sco red from 210 predatory attacks revealed that ,for both successful and un successful attacks, Polaris Breach was the most commonly employed initial stri ke,while Surface Lunge was the most frequent second event, closely followed by Lateral Snap. Examination ofv ideo footage, still images, and tooth impressions in decoys indicated that white sharks at Seal Island biteprey obliquely using their anterolateral teeth via a sudden lateral snap of the jaws and not perpendi cu larlywith their anterior teeth, as previously supposed. Analysis of white shark upper tooth morphology andspacing suggest the reversed intermediate teeth of white sharks occur at the strongest part of th e jaw andproduce the largest wound. ,,,Ihite shark predatory success at Seal Island is greatest (55% ) within one hourof sunrise and decreases rapidly with. increasing ambient light; the sharks cease active predation on sea lswhen success rate drops to 40%; this is t he first evidence of cessation of foraging at unproductive times byany _predatory fish. At Seal Island, ",,-hite shark predatory success is significantly lower at locations wherefreque-Flcy of predation is highest, suggesting that white sharks may launch suboptimal strikes in areas ofJSreatest intraspecific competition; this is the first evidence of social influence on predation in any elasmo branch. Idiosyncratic predatory behaviours and elevated success rates of known individual white sharks atSeal Island suggest some dcgrcc of trial-and-error learning. A hypothetical decision tree is proposed tha tmodels predatory behaviour of white sharks attacking Cape fur seals at the surface.

INTRODUCTIONPredation is a ta ct ica lly fluid event, the outcome of

which depends upon the behaviour of both prey andpredator. Behaviour imposes costs in terms of energy,time, and risks that must be balanced against survivalbenefits, such as resources nceded for self maintenanceand reproduction (Ellis, 1986). Since a prey animal haseverything to lose in a predation event, one would expectit to be willing to commit any amount of energy towardescape. Conversely, if the energetic cost of a predationattempt is too high or the likelihood of capture too low,one would expect a predator to abandon the attempt(Bennett, 1986). Foraging models comprise three elements:(1) decisions made by a predator to attack or not attack aprospective prey; (2) currency, such as energy costs orgains ; and (3) co nstraints, or the factors defining the relationship between decision and currency (Gerking, 1994).Behaviours reflecting predatory choices can be better

understood by dividing predation events into componentparts. The predation sequence is usually divided into fivestages: detection, ident ifica t ion , approach, subjugation,and consumption (Endler, 1986). Optimal foragingtheory (OFT ) predicts that a predator should exploit theprey type most energetically advantageous in terms of netenergy content minus search and handling costs (Gerking,1994). The OF T further predicts that predators should beselective when high-qu ality food is abundant (Helfman etaL, 1997). Learning appears to be the underlyingmechanism for adaptive behavioural responses in fish foraging (Dill, 1983). Th e high level of development of th eelasmobranch brain (Northcutt, 1977, 1978) implies thatsharks ought to be capable of the types of foraging decisions predicted by Dill (Bres, 1993).

Natural predation by sharks is seldom observed in thewild, due in la rge part to their high vagility and thevisually concealing nature of the marine environm ent(Myrberg, 1987; Klimley et aL, 1992; Bres, 1993). Th e

Journal oj he Marine Biological Association oj the United Kingdom (2005)
http:///reader/full/elasmo-research.orghttp:///reader/full/elasmo-research.org


2/15

R.A. :'1artin et al. behaviour

tive with which it ca n be observed an dupon pinnipeds at the

Despite thisknown of lhe predatory behaviours the 'white shark

under natural conditionsmost what is known aboutprey capture and

no t:vidence to support thisan ethogr;ml for white sharks

the Huallonprecluded

strike.A white shark's

exhibit atypical lamnoid dental pattern, with the upper dentition

'featuring markedShimada,

how white sharksprey capture has 110t been

themselves intopartly on

haul-out sites, whitesharks often aggregate after another has made a kill an dseveral discrete of social interactions havc bcendefined (Klimley et al., 'HtilJournaL {if the Alarinc Bid(;gica! oj" Ille Unifed Kingdom (2005)

sharksdecisions that

presented. A whiteincluding four

an d event sequencethe

effects on white sharkA decision tree model-

ling behaviour of white sharksfur seals at the surface is proposed.

LLL ' - . L rU . JU ANDbehaviour ofwas studied bC'tween

Africa. Seal Island isfoot of False Bay, with

north-southmeasures elevation IS7 m above the and is centred atmately 34)8'6"S 1834'OO"E. Th e underwaterof Seal Island features amost of the western side

shallow shelf north-cast side. Th e watersSeal Island were divided into six sectors,

Seal Island

and return toun published

point, termed theoutcrop located off thethe juncture of Sectors

fur seals occurs in

or lr 1were made at every

weather permitting, of two to eight froma 8-m outboard boat from 1997 to 2000, an d byteams of four to 16 from two such from 2001 to 2003.At least part of the research team some 200 daysper on the water. Arrival at Seal Island was before

at 0730 h, sea conditions permitting. By
http:///reader/full/LLL'-.LrU.JUhttp:///reader/full/LLL'-.LrU.JU


3/15

Predatory behaviouT ifwhite sharks R.A. Martin et al. 1123

A

B Tableo Bay

Seallsland

Cape Point

Figure 1. Study site: (A) location of False Bay, South Africa,with the 200 m depth contour indicaled; (B) loca tion of SealIsland wi thin Fal se Bay, showing the main path ofpinnipedmovement leaving and returning to th e island (grey triangle);and (C) depth co ntours (m) and sec tors (circled numbers) ofarea around Seal Island, showing a typical decoy to\v-pa th(grey) along the drop-off on the south-west side of th e island .

stationing at a terminus of Seal Island, a single vesselcould sur vey some 270 0 uninterrupted to a distance of

~ 3.5 km. With two vessels at opposite termina ls, nearlyall the water surrounding Seal Island could be surveyedto a co mparable di stance.

Sea l group size was est imated as solit ary, 2- 4, 5-10, or11+, an d their di rec tion of travel relative" to the islandcategorized as either outgoing from or incoming. Capefur seals divided into four classes that' combine theirlength , m a t ~ r i t y state, and sex (when determinable underfield = nditions): Class I: neonates (black pelt, 2 m).

Surface predatory events were detected by one or moreof the following: (I) white sha rk breach, with or without asea l in its mouth; (2) a sudden change in the travel behaviour of seals, switching from directional porpoisi ng toeither zigzag evasive manoeuvres or head-stand subsurfacesca nning, with indications of a shark in pursuit (largesurface boils and/or direct observation of the shark); (3) alarge splash accom panied by a blood stain, oil slick, and adistinctive odour, sometim es accompanied by secondaryindicators such as a floating sea l head , exciscd heart and/or lungs, an d entrails either floating on the surface ortrailing from the gi ll openings of a white shark in theimmediate vicinity; (4) highly localized c ircling an d /o rplunge-diving black-backed kelp gu lls (Lams dominicanisvetula) a nd other seabird s, with kills often accompaniedby active competition for sea l entrails by seab irds.Sequences of predatory events were scored and documented using standard techniques. Circumstantialevidence of 10 subsurface strikes by white sharks onCape fu r seals were recorded during the study period, bu tcould not be scored due to limited visibility.

1.40%

1.00%

0.60%0.40%0.20% ,+------- - - r

B ~~ L 1

cA2 Al S Y m ~ h ~ ~ ~ , ~ ,

L1

\ ( ~ , \Al.Ai" ~ bInterspace . \ a

I n t ~ ; ~ ; a c e \ DL2 L 1 L2 Interspace ~

D

.'.;

0:wuCo

E

O.OQ ()/o . , .-- 1- - ->-- -1 -1-T --1 I ---+--+--,;.----1L2L 11L ZAI lAZA lAZA 2A1 /-1L 1L2L

Left Righi

Figure 2. White shark denlition and terminology: (A) labial(inner) view of anterior part of a representative upper jaw ,showing location of the intermedial e bar between the intermediate (I) and first lateral (L I ) teeth; (B) distal v iew of I andme sia l view of L I, showing curved cutting edge of the former;(C) jaw terminology, tooth identification and measurements,with position of the int ermediate bar indicated by chord a- b;and (D) Dice diagram of interspace ratio between successivepairs of upp er teelh, where vertical bar =range, horizontalbar=mean, white box=standard deviation, and hashedbox=95% confidence limits. In both (C) and (D), the verticaldashed line in di cates head axis through the jaw symp hysis.

Every effort was made to minimize observer effects onpredatory events. Predatory events were classified as eithersuccessful, in which the seal wa s consumed, or un success ful , in which the seal escaped. Duration of predatoryevents wa s timed Ii'om initial st rike to either consumptionor the prey or continuat ion of seal travel on an identifiableco urse fo r ? 30 s; in those few cases in which a sea l wa skilled but not co nsumed (N =10 ), end of the predat ionevent wa s considered the time after which the floatingcarcass was unattended by a shar k for ~ 120 s. Di stancc ofpredatory events from the islan d was estimated to thenearest 10 m with the a id of an on-board GlobalPositioning System. Shark length was estimated to th enearest 0.1 m by comparison against known dimensions ofthe attending observation vessel. Shark sex was based onthe presence or absence of claspe rs, verified wheneverpossible by polecamera im ages of the cloaca from below.

Journal of he lv/mine Biological Association of he United Kingdom (2005)
http:///reader/full/1.l-1.5mhttp:///reader/full/1.l-1.5mhttp:///reader/full/1.l-1.5mhttp:///reader/full/1.l-1.5m


4/15

11 24 R.A. M artin et al. Predatory behaviour i fwhite sharks

~ . - - - - - - - - - - - - - - - - - - - - - - - - - - - 300B Outgoing, n !II 371::t 100 '"- 250::l (J Incoming, n 267 e(!) 80 , 200.... Io ..IIIII 60 '0 150'0 40 I.c " 100E ::l20 Z 50

o 0731 - 0831 - 0931 - 1031 - 1131 - 1231 0830 0930 1030 1130 1230 1330 2 4 6

Outgoing, n 352r- o Incoming. n ,:. 314--

t - . . .JIITlmo (hrs) Seal Island Sectorc D

Successful) n = 166:: r ----------- 16o Unsuccessful, n :J 124 14120 -if---------- - ---------------1

.!ii 12~ - - - - - - - - - - - - - - - - - - - - - - - - - tI 10w '.---------- - - ---------- --1 ..:1 860 4.----------------------------------14~ ~ - - - - - - - - - - - - - - - - - - - - - - - - - - - ~ - - - -220

o ' tr..u,.IJ.il.JU1, , '". I n 11 I] n fl

, 2 3 4Duration of Attack (minutes) Breaches

~ - - - - - - - - - - - - - - - - - - - - - - - - - - - - ~ Succ6'ssful, n 20a Unsuccoliisful, n :II; 16

E 450 - F 700 ... -400

.. ----r- -r

-, Soccessful, n =.. 9.81o Unsuccessful, n 12 1043

r- r _ .

+- I-r =I fill-, _ ,-r .. - ~ - Succ itoi. n =1010600 r350 o Un suec.. .fui. n =1076500300tl.." 250200 ..: 300150

100 200 r-50 100 0 ~ - - ---=--. '- r-- C I l ~-200 201- 601- 801- 1001- 1201- 1401- 1601- 1S01

4,90 600 800 1000 1200 1400 1600 ~ ~ 8 0 0 2000 2 3 4 5 6Distance from Seal Island (m) Seal Island Sector

Figure 3. CApe fur sea l movements a nd whi te shark at tac k patterns at Seal I sland, Sou th Africa: (A) number ofseal groups versusti me; (B) number of sea l g roups versus I sland Sector; (C) number of a tt acks versus duration of attack; (D ) number of attacks versusnumber of breaches; (E) number of attacks versus distance from Sea l I sland; and (F) number of attac ks versus Islan d Sec tor.

Ind ividual whi te sharks were identi fied at predatoryevents whenever po ss ible. From 199 7 to 2000, 73 colourcoded stre amer tags a nd 20 pinger tags were attached towh ite sharks at Seal Island. Tagging of white sha rks inwestern and southern Ca pe waters was banned byM arine and Coastal M anagement in 2000, resulting inin itiatio n in 2001 of a non-invasive identification techniquebased on cataloguing an d coding individual white sharksby natural pigmentation patterns and other per sistentma rk s; thi s technique will be described in detail elsewhere.Identifica tion of individua l white sharks relied on fieldrec ords of shark length and sex (if known), tag pl ace ment,type, and (where relevan t) co lour cod e, plus naturalpigmenta tion pattern.

Data on white shar k at tack frequency and success withrespect to sea l group size, di rect ion of travel, and age class, time of day, distance from shore and by Islan dSector were compared via one and two-way ana lysis ofva riance (ANOYA) and Tukey - Kramer tests (Type Ierror = 0.05) with replication . Shark length and Cape fursea l size-class were easier to identify in successful attacks,Journal of the Ma rine Biological Association of the Un ited Kingdom (2005)

thus one-way ANOYA was employed in these cases to minimize th e effects of bias in the data . Frequency and eventsequence ana lyses of behavioura l units of our predatoryethogram fo llow the methods of Klimley et al. (1996a).

After frequency of predatory ac tivity attenuated eachda y, some 2-2.5 h after sunrise, a seal-shaped decoy wastowed behind our research vessel to elicit strikes by whitesharks . D ecoys were ca rved from compressed closed-ce llrubber camping mats that preserved individual toothimpressi ons and did not injure attacking sharks . Th edecoys were towed th rough Sectors 3, 4, and 5 of Sea lIsland over the 15-18-m depth contour some 8+ m behindthe vessel a t a constant speed of 2.5 km h - I (Fig ure lC) .St rikes agai nst decoys were limited to < 3 per day, a fterward the island was circumnaviga ted slowly at a distanceof 4 to 15 m, scanning for shark bitten Cape fur sea ls a nddocumenting wound characteristics (location, relat iveseverity, degree of hea ling).

Str ikes on decoys were documented via videographya nd still photography. Ja w position of individual uppertooth impressions could readily be identified following
http:///reader/full/tr..u,.IJhttp:///reader/full/tr..u,.IJ


5/15

the terminology of Shimada (2002). Interspace measurements were made following Collier (1992, 2003) from 25white shark jaws , uniformly prepared by G. Hubbell ,representing individu a ls of both sexes an d total lengths(TLs) ranging from 158 to 594 cm. To correct for sizedifferences among sharks sampled, interspace measurements were converted to per cent TL, resulting in a newmeasurement here termed ' interspace ratio'. In terspaceratios of th e first five upper an terior teeth on either side ofeach jaw were compared statistically. Tooth identificationand representative interspace measurements are indicatedin Figure 2A&C. For comparison , 44 cleaned and driedwhite shark ja ws were examined from the reference co llections of the Los Angeles County JvIuseum of Na turalHi story (L AC M ), British Museum of Natural History(BMNH ), the Cape Town Museum (CTM ), the NatalSharks Board (NSB ), and the private collection ofG. Hubbell (Jaws In ternational ).

RESULTSA total of 2088 natural preqatory int erac tions between

white sharks (Carcizarodon carcharias) and Cape fur seals(Arctocephalus pusillu; jJltSillus ) was do cum ented at thestudy site between 1997 and 2003. Most predatory activityocc urred between late May an d late August, with theg reatest frequency occurring bet \ ;een mid-Jun e and midAugust (winter). Up to 25 pr edatory interactions wereobserved in a single day, with a mean of 5.6 attacks perday. Mean predatory success rate was 47.3%. lVIovementof sea l groups away from and toward Seal Island wassignificantly greater in the early morning between 07 30and 0930 h (P < 0.0001 ), but rema ined high until at leas tea rly afternoon (Figure 3A), with most movementcentred on the Launch Pad at the junction of Sectors 3an d 4 and significantly more seals ~ i n g from orre turning to Sector 4 than Sector 3 (P < 0.0001) (Figure3B). Group size of sea ls attacked ranged from I to at least15, with frequency and success rate decreasing withincreasing group size an d significantly more attacks onsolitary seals than any other group size category (N =973;P< 0.0001). Both incom ing and outgoing sea ls wereattacked, but frequency (N = 287, P < 0.0001) and successrate (N =287, P < 0.05) were significant ly higher onincoming seals. Sea ls representing a ll size-classes exceptClass 1 were attacked , bu t predatory frequency decreasedwith increasing sea l class and was significantly higher onClass 2 sea ls than any other size-class (N =1088,P


6/15

1126 R.A. Martin et a1. Predatory behaviour ofwhite sharks

Figure 6. Predatory behaviours in white sharks at Sea l Island, South Africa: (A) Polaris breach; (B) surface broach; (C) la teralbroach; (D) invcrtcd broach; (E) surface lunge; (F ) surface intercept above water; (G ) lateral roll; (H ) surface arc; (I ) lateral snap;(J) direct surface app roach; (K ) killing bite; (L ) surface grasp horizontal approach; (M) subsurface carry; (N) surface feed; and(0 ) lateral head shake.

of attacks ranged from < I min to 27 min, although nearly2/3 lasted ~ 2 min (N = 290, mean=3.0, SD=3.73). Therewas an inverse relationship between attack duration an dsuccess rate as well as between number of breaches andsuccess rate (Figure 3C,D ). Mean duratiOn of 166successful attacks was 1.42 min with SD=2.1S min ; if th e27-min outlier is discounted, mean duration of successfulattacks was 1.27 min with a SD of only 0.S8 min.

Upper teeth of all prepared wbite sbark ja ws examinedfeatured a reversed intermediate tootb, located ju stanterior to the intermediate bar ncar the point ofmaximum curvature of the upper jaw (Figure 2A ). Theblade of this tooth, I , typically has a curved Lo slightlysinusoidal distal cutting edge, unlike the nearly straightmesial cutting edge of the 1st lateral tooth, Ll (Figure2B). Interspace ratios of 1- Ll of left and right sides of the

.JournaL 0/ [he Marine BiologicaL Ilssociation q[lhe United Kingdom (2(}(J5)


7/15

jaw were significantly larger than for any other tooth-pairon either side ofthejaw (Figure 2D).

Over the study period, 121 strikes on towed seal-shapeddecoys were documented. Identification of individualupper tooth punctures in compressed rubber decoys indi-cated that all strikes were asymmetrical an d inconsistentwith being grasped by the shark with the anterior teethwhile oriented perpendicular to the longitudinal (travel) .axis of the decoy (Figure 4). Videotape footage and stillimages revealed that decoys are typically graspedbetween the anterolateral or lateral teeth via a sudden,lateral snap of the jaws (Figure 5A). This is consistentwith footage and still images of initial (capture) bites onCape fur seals (Figure 6F & I) and wbunds on escapedseals (Figure 5B).

Based on observations of natural predations an d strikeson seal-shaped decoys, a predatory ethogram consisting of20 behavioural units was constructed. These units ar edivided into four functional phases and, in approximatesequential order, defined as follows.Phase 1: Initial strike

Polaris breach (POL): the shark leapt partially orcompletely out of the water in a vertical or nearly verticalhead-up orientation, with or without a seal grasped in itsjaws. In POL that launch a shark completely from thewater, the shark may clear the s'{lrfacc by as much as 3 man d often rotates tail over head around the centre ofgravity, located lj3 along its standard length, re-entering the sea head-first close to its original exit point.During POL in which the seal is not grasped by the sharkan d continues evasive manoeuvring, the'shark often turnsits head mid-flight as though visually t r a c k i n ~ the seal'smovements (Figure 6A). .

Surface broach (BRS): the shark leapt partially orcompletely 9ut of the water in an upright ;'rientation withits body aXIs forming an angle with the horizon'between45 arrd 0 degrees, with or without a seal in its jaws. InBRS that launch a shark completely from the water, theshark usually clears the surface by :( 1m, re-entering thesea in the direction of travel 0.5 to 1.5 body lengths(BLs) away from its original exit point and ,


8/15

1128 R .A. M artin et al. Predatol)1 behaviour of wh ite sharks

sight. The RO L occurred underwater or partially abovethe surface with a pectoral fin proj ecting in to the air(Fig ure 6G).

Surface arc (ARC ): the shark swam slowly to moderately qui ckly at or near the surface adop ti ng a broad, semicircular co urse around a surface-borne sea l. During ARC,the diameter of th e ar c descr ibed is 2- 3 BLs, the sharkswam between 0.5 a nd 2 BL S - I, an d its eyes rotated intheir sockets, ap parently to ke ep the st ricken sea l in sight(Figure 6H).Phase 3: Prl!)l capture

Lateral snap (SNL): the shark captured a rapidlyfleeing pinniped via a sudden latera l snap:of the jaws,pro truding the upper jaw an d grasping the prey with itsanterolateral teeth. During SNL, the sha rk's head isflexed sideways some 10 to 15 toward the prey. The SNLis ver y rap id and usually lasts < 0.5 s. Upon contact withthe prey item, the eyes are rolled tailward in th eir sockets,exposing the whitish sclerat ic coat , after which the head,with or without the seal grasped i)1 the jaws, is swu ngmedially until co llinea r with th e pody ax is (Figure 61).Often followed by REP.Repurchase (REP ): the shark shifted its per ipheralgrasp on a seal, leaving th e lowe r dentition inserted whilerapidly lifting the snout, the reby removing the upperdentition, then qui ckly p rotruding the upper jaw andbringing the up per teeth back in to co ntact with the preyfar ther from the periphery than previously; often accompanied by a sudden lateral movement of the head.

Killing bite (KIL): th e shar k delivered a deliberate,powerful bite w ith the an terior teeth to th e hea d an d neck-- of a captured seal, apparently to kill i t pr ior toDuring KIL, up on initi a l contact with the prey,- th e eyesare rolled tailward in their sockets, exposing the whi tishscleratic coat. Often followed by LH S (Figure 6K ).Direct surtace approach (DIR): the sharkperformed a relatively slow, or iented approach along th esur face to a surface-borne dead or severely injured butweakly swimming seal. During DIR , the shark's swimming speed was typica lly ~ I BL s - I (Figure 6J ). Usuallyfollowed by FD S or FDU.

Surface grasp, horizontal approach (GRH): theshark slowly and deliberately approached a surface -bornedead or ot herwise incapac itated (non-swimming) sealalong the surface and grasped it with the anterior teeth .During GR H, th e shark 's swimming speed was usua lly< 0.5 BL S- l (Fi gure 6L). Usually the first reacquisitionof a food item after REL.

Surface grasp, vertical approach (GRV): thesha rk slowl y and deliberately approached a surface-bornedead or otherwise incapacitated seal from below at anan gle of 45 to nearly 90 0 and grasped it with the anteri orteeth. During GRV, th e shark's swimming speed wasusually < 0.5 BL S-lPhase 4: Feeding

Subsurface carry (CAR): the shark slowly ca rried adead or otherwise inc apacita ted seal underwa ter for> 3 s,transpor ting it > 6 m before feeding. During CAR, th eshark's swimming speed wa s reduced to 0.5 BL S- I a nda mplitude of each caudal stroke "vas increased nea rly50 % over that ex hibited during normal swimmingJou rnal of the lv/arine Biological AssocialioTi of he Un ited Kingdom (2005)

70

so40

2010

7060

2010

151 Event I , ,liuccessful, n = 65o Unsuccessful, n = 145---

I J, !Ifr= 1

Successful , n = 65 2nd E v e n --Unsuccessful, n ;;'-1f ; Z - -f - .

I, IIIIi -.n1L. ~ ~ . ~ r ' - - -

70

50 + -'E.. 40

302010

-'0"(I)a:'"

-'0:'" ii'IIIz::>-'

;;; ::>;;; ...J00: u0: - 'Z(I)7060so40

Q)" 30

10

-'0"(I)0:III

-'0:'" iii'OJz::>-' ::> -'00: u0: - 'Z(I)

70

- - - ~ - - - - - - - - - - - - - - -3,d Event.0 + - - - - ~ - - - - - Successful, n =65D unsucce:.::s "u-'-'-,,-n""="'8""0----I-- --- - -'

- ~ l- '

70 -. ----,-- - - - - - 60 +-____h""E'-v_e_n_t........__-: -_ 50 ---- -- -

"- -' DC X > a: ::> (I) -'UJ 0:i: 0: 0: a u "". u." '5 ::!" "

0: (I) (I)"- -' ::> -'UJ UJi: '" 1!2 1:: 0: a u u. '5" '""

- - - - - - - - - ---, Successful n = 55 ------OUnsliccessful, n = 4

L--

41h EventI

- I11 --_.1 11 n,t l '

Successful, n = 64o Unsuccessful, n :; 33-

--It , I I

60so

C 40..

51h E

it 30+ - - - -

nt Successful n = 62o Unsuccessful, n = 12

20 +-- - - -__ j1 - --- ----- - -_1---1

u -'0: Z '"

'E 40 +---- - - 1f----- ' - - - --.u - - - - - _ _ j ~ - - - - - - - - - - - - - - - - - - -Q.~ ~ ~ , ~ - ~ - - ~ ~ - . ~ ~~ ~ - ,

-' -' iii' z ::> -' u -' "- -' 0: X > 0: (I) ::> -'0 '": 0: OJ ::> z z 0 0: Z UJ i: 0: 0: ;I:'" UJ"- en -' - - 0: (I) 0: a u u." -' 0:'" J \ J " "1\Initial Secondary Prey FeedingStrike Pursuit Capture

Figure 7. Pe r cen t occurrence of white shark behavioural unitsscored du rin g firs t six events in successful versus un successfulpr ed atory at tacks at Sea l I sland , South Africa (n total=210at tacks) .

I


9/15

Predatory behaviour of white sharks R .A. Martin et al. 1129

Table 2. Pr oba bilities that each behaviour unit precedes orfoLLows each oth er behaviour uni t during predatory attacks on Cap efuT sealsat Seal Isla nd , South Africa (N = 210 attacks) .Subsequent behaviour

Initial strike Secondary pursuit Prey capture FeedingPOL BRS BRL BRI LUN INA INU ROL ARC SNL REP KIL DIR GRH GRY CAR LHS FDS FDU REL

Code N II 29 0 0 152 5 2 6 12 110 36 51 7 14 8 56 53 II 53 49POL 103BRS 88BRL 13BRI ILU N 99I NA 5

"" I NU 20 RO L 6ARC ]( )..cQ,) SN L 100~

bJl RE P 36= KI L 51~uQ,) DIR 7Q,)

GRH 14""o< GRY 8CA R 54LH S 52FD S 0FD U 3RE L 13

no,0.0:'>

ft. .(U()

0.380.560. 38

0.0 I0.0 10.03

0.07 ' 0.44 0.05 0.0 I 0.04 0.07

0.500. 500.07

0.17010 0.10

004\ 0.010.25O!H U.ORl.000.29 0.01 0 .010.80 0.200.50 0.500.17 0.170.20 0.10

0.35 n.:n 1l.0 I 0.02 0.01 0.21D.H a.(If\ 0.06 om 0.03 0.33Dll6 0.63 0.12 0.04 0.03 0.080.1 0.29 0.14 0.14 0.14 0.14

(Ul7 0.50 0.07 0.14 0.210.63 0.13 O. \3 0.130.02 0.78 0.02 0.13 0.06

D. D4 0.1 0 0.06 0.69 0.1 20. 33 0.33 0. 330.62 0.38

(Figure 6M ). Often, at leas t one other shark was visible inthe immediate vicinity a t the onset of CAR.

Lateral head shake (LHS): the shark grasped a deadseal in its m outh and shook its head violently from side-toside, the 5nout describing an arc of 9Q,, removing apie ce from Jh e carcass (the remains of which typically

Abandon

floa ted to the sur face). Period of each LHS was 2 s(Figure 60 ). Us ually followed by FDU.Surface feed (FDS): the shark consumed a dead sea l

at the surface, usually in on e or two bites, circling slowlybut tightly (turning radiu s 2 BL) between bites . DuringFDS, circling speed was < 1.5 BL S - I (Figure 6N). Th e

Figure 8. Hypothesized decision tr ee of preda tory tac tics employed by whi te sharks on surface-borne Cape fur seals at Seal Island ,South Africa. Three-le tter codes for individu al behaviour units ma tch those used in the e th ogram (see text for d etail s) .Jo urnal of he lvIanne Biological Association of lze United Kingdom (2005)


10/15

1130 R,A, l'vlartin et a1. behaviour sharks

orreleased a whole or

which floated to the surface, and didnot reclaim it for 10+ min, least one other shark

Th e seal carcass

and event sequence of thevioural units wcre scored from video

on 210 of which 65 were ~ L l \ ~ L C , ~ ~ la ~ ~ d

unitsattaeks consisted ofunits For both successful and unsuc

cessful the mostinitial

than smaller individuals and ha d asucccss rate than sharks of the sameto Seal vs 58.8%,

kills were attended by abut are sometimes attended by up to four

at least 2B events documentedthe shark that a seal kill

was nol the same individual that consumed the carcass.on 10 2002 at 0928 h, 3.5-m white

2 fu r sealbefore it consume the moribund

it was followed at dislance of 2 BL and withineonsurned it; thethe seal was not

Ii rhited energy stores,Class 2 scals may also naive and thus easierfor white sharks to ambush and Loneseals are unable to share or defensive dutie s and

white sharks thangroups. seals

groups and may be influenced torporthus be more vulnerable and less attentive than

IS

Jt IS unknown howspeed, Between 94muscle mass is

whieh

manoeuvres whentoward the shark'ssharkThe anumerous capture attemptslower its chances of

This that once a shark has launched itsinitial whereaboutsfavour the seaL Thus, like

fish, white sharksIsland arc ambush

Most white sharkIsland occur

strategy window and has limitedWhite sharks seal from the

appear to select bottom can below it beforeJoumalq{ Ihe Marine Biological n,uucwwm Qllhe United Kingdom (2005)


11/15

would

white sharkshave a visual and tacticalthe surface.

It is significant that white sharks at Scal Island appearto limit active predation to within a few hours""'""",,,!', sharks in theforage at high densities of calanoidconcentrated at thermal fronts but swirn out of theseareas on straight courses whenfell below ! g (Sims &

1998). Similarly, our data showat Seal Island cease activewhen success rateevidence of cessation of

fish.

Seal Islandappear resident in Falsefrom Seal Island and closer

I I I Falseto prcy, which may be more

to capture. This idca is an dbut remains to be tested.

introduccdlamnoid toothshark

which separatestooth pockets. In this

Carcharias and u{,wnWjfJlSteeth and the so-called

in his

IS here.upper dentition all modern white sharks features

intermediate teeth. If these teeth have acannot in isolation bu t in

with the teeth that it. The teeth of adolescent andadult white sharks among laml10ids in having

serrated crowns and in lateral teeth thatbroader than those of any other

The first an d second anterior}ourno.l q/thr krafine Biological Iissocialior! q/I!t


12/15

R.A. :vrartin et a1. behaviour sharks

female 1Nhite shark that has been recorded atSeal Island since 21 1997, initiated a surface attack ona Class 2 fur seal 300 m offShore in Sector 3. crhe

began with a Lateral Broach at 0813followed 6 later a Surface Broachlater by a Surface Lunge the seal divedpeared from view for 15to locate thefall back into the

to function to increaseof a white shark's grasp on

4, is clear that REP notaXIs to nprn f ' nn

cular to the long axis of the shift in graspa sudden lateral movementof the head while the lower dentition remainedinserted into the crhe lower dentition of whitesharks is less flattened and blade-likethat of the upper

Anterior lower tooth ventralsurface of the are less dist inct the upper

wi th shredded consistentTh e REP can be

of prey.may reduce risk of toseal or it simply assert

prevent another white shark26 individual white haveIsland and as many as

'four observed a kill. amongwhite sharks for a disabled seal at Seal Island is likely

i n t e n s e ~ and CAR may hElp preventanother white shark. Like et al.found no evidence to support the

The brief duration of successful white sharkevents at Seal Island may reflect the

relatively small mean size of prey theU) l l l l f lVUUUl l among white sharks at Seal Island.

Lateral head shakeof a white shark's

Performed

further increases theThe LHS

are action patternssharks

grey reef sharkmay ritualized to contexts socialsignal. Such an was observed on 25

2000 at 1325 h, w h e n ~ m a baited context and

of about ,4 m ~ a 4-m male white shark3 s toward a 3.5-m white shark of

indeterminate sex, which acceler ated away anddid not return.

Food release may represent food dueunder social threat. On

4- al 0735 h, a

and/orunknown.

Surface-based and the concentrationnear the Launch Pad afforded advan

an d limitations. 'Thousands of whiteattacks on

4-m fromorientation

vations limited to behaviou rs visible from surfacerestricts the present study to the last three stages of

andviour units to be measured is oneand difficult decisions 10 be made.

and event sequence of behavioural units aresimilar to those Klimley etbased on white shark lions andother otariids at SFI, theysuccessful unsuccessful attacks.and orientation of the initial strike of an attacking whiteshark may be rather reflectunder of tactical conditions.

a vertical (GRV)ma y reflect individual

or presence of a compeneeded to addressof

scored.of predatory units

fur seals at Sealbehaviour is

and

as a decision treeupon via up to five behavioural units suggests

}ournai rt/lhe Afarine Biological flssm;WSWIl Q,F thf [!! l i lat (2005)


13/15


14/15


15/15

Documents

Predatory Behavior of White Sharks