60(3), 1998 9

Introduction

Like the majority of otter trawls,prawn trawls typically are poorly selec-tive fishing gears, and in addition to thetargeted species they also catch andretain large quantities of nontarget spe-cies (termed by-catch), which oftencomprises a diverse assemblage of smallfish, crustaceans, and cephalopods (forreviews see Andrew and Pepperell,1992; Kennelly, 1995). While some ofthis by-catch may be retained and soldcommercially (Broadhurst and Ken-nelly, 1997), large quantities are oftendiscarded at sea.

Several studies have examined thefate of discarded by-catch from prawn

Bottlenose Dolphins, Tursiops truncatus, RemovingBy-catch from Prawn-trawl Codends During

Fishing in New South Wales, Australia

M. K. BROADHURST

The author is with the N.S.W. Fisheries ResearchInstitute, P.O. Box 21, Cronulla, NSW 2230,Australia. His present address is UniversidadeFederal Rural de Pernambuco-UFRPE, Departa-mento de Pesca, Laboratório de OceanografiaPesqueira, Av. Dom Manuel de Medeiros, s/n, DoisIrmãos, Recife-PE, Brazil, CEP: 52.171-900.

ABSTRACT—During a fishing trip torecord video footage of fish escaping froma by-catch reducing device located in a com-mercial prawn trawl, two bottlenose dol-phins, Tursiops truncatus, were observed toactively manipulate the codend at the sea-bed, removing and consuming componentsof catch (mostly juvenile whiting, Sillagospp.). The observed feeding pattern suggestsa well established behavioral response totrawling activities and is discussed withrespect to (1) the potential nutritional ben-efit that dolphins may derive from such ac-tivities and (2) the effects that scavengingmay have on selectivity of the gear.

trawling and show that in some cases itmay contribute significantly to the di-ets of various scavenging predators, in-cluding seabirds (e.g. Phalacrocoraxvarius, P. melanoleucos, Anous stolidus,Sterna bergii, and S. hirundo) (Blaber andWassenburg, 1989; Hill and Wassenburg,1990; Blaber et al., 1995), crustaceans(e.g. Portunus pelagicus) (Wassenburgand Hill, 1987; 1990), fish (e.g. Nemip-terus spp. and Pentapodus spp.), sharks(Carcharhinus spp.) (Hill and Wassen-burg, 1990), and dolphins (e.g. Tursiopstruncatus) (Leatherwood, 1975; Cork-eron et al., 1990; Hill and Wassenburg,1990). Many of these studies have usedeither visual census or underwater cam-era and video, at the surface and on thesea bed, to document the feeding be-havior of various predators scavengingby-catch discarded from prawn trawl-ers. There is also some anecdotal evi-dence to suggest that some predators,such as bottlenose dolphins, Tursiopstruncatus, may feed at prawn trawls dur-ing towing (Leatherwood, 1975), how-ever, there is an absence of informationand visual evidence on the extent towhich these sorts of predators interactwith the trawl. This communication docu-ments one such interaction by providing thefirst videographic evidence of dolphins at thesea bed actively manipulating the codendof a prawn trawl to remove by-catch.

Materials and Methods

These observations were made atnight onboard a commercial prawntrawler (13.8 m) fishing on groundsnortheast of Yamba, New South Wales

(NSW), Australia (Fig. 1) in Septem-ber 1996 during a trip to record film offish escaping from a composite square-mesh panel (a by-catch reducing devicedeveloped for NSW oceanic prawn-trawls—Broadhurst and Kennelly,1996; 1997). Three Florida flyers (meshsize 42 mm) each with a headline lengthof 12.8 m were rigged in a triple gearconfiguration and towed at 2.5 knots indepths ranging from 18 m to 22 m. Thestarboard outside net was fitted with acodend containing a composite square-mesh panel (Fig. 2). A “Photosea co-bra” underwater video camera wasmounted over the anterior section of thecomposite square-mesh panel, facing aftalong the top of the codend (Fig. 2). Thiscamera was linked, via coaxial cable,to a control console and PAL videomonitor onboard the vessel and suppliedwith 240 volts of electricity. Two 240watt submersible lights, facing aft, weremounted anterior to the camera on ei-ther side of the codend (Fig. 2). Theselights were necessary since the camerawas used at night with zero visibility.The nets were set and towed accordingto normal commercial operations withthe camera switched on immediatelyafter the gear touched the sea bed andthen turned off again prior to hauling.

Results and Discussion

During three successive 60–90 mintows during the night, approximately 10–20 min after the camera was switched onand shortly after the lights were activated(providing a visibility of about 1.5–2 m), two bottlenose dolphins were ob-

10 Marine Fisheries Review

Figure 1.—New South Wales coast and location of area trawled.

served to be in the process of slowlyapproaching the posterior end of thecodend (Fig. 2 and 3A). The dolphinsswam directly into the rear of thecodend and using their rostrums andforeheads, pushed the diamond-shapedmeshes forward and up with consider-

able force (Fig. 3B). This movementdisplaced the catch forward, effectivelyincreased the fractional mesh openingsin the codend, and resulted in the re-lease of large numbers of small organ-isms and fish, mostly juvenile whiting(Sillago spp.). The dolphins retreated

and consumed some of these fish as theydrifted from the codend (Fig. 3C) andalso, by tilting their heads laterally,swam forwards and actively removedother individuals that were trapped be-tween meshes. This behavior was re-peated for up to 20 minutes during eachtow. The dolphins were not observed tochase nor consume any of the live whit-ing that were escaping from the com-posite-square mesh panel. These fishtended to rise up above the dolphins andwere quickly lost from view.

To determine if the lights attached tothe camera had any contributing effecttowards behavior, they were switchedoff (resulting in zero visibility), for upto 45 seconds, and then on again. Thiswas repeated 7 times while the dolphinswere at the codend, but had absolutelyno effect on their feeding pattern, indi-cating that their behavior was well es-tablished. For example, at the exactmoment the lights were switched on, thetwo dolphins were observed to be ei-ther in the process of (i) actively remov-ing fish from the codend, (ii) forcingthe codend forwards and upwards, or(iii) consuming fish that were releasedas a result of the previous action. Fur-ther, on two occasions when there wereno dolphins at the codend, the lightswere switched off for approx. 5 minsand then on again, revealing two dol-phins displaying the same routine be-havior as that discussed above.

Because it was necessary to turn thecamera off prior to hauling the trawls,it was not possible to determine if anydolphins followed the codends to thesurface. However, at the end of each towup to 5 bottlenose dolphins were ob-served foraging around the vessel andat the trawl while the codends and cam-era were brought aboard. These indi-viduals remained in close proximitywhile the catch was sorted, disappear-ing only when the vessel was underwayand setting of the trawls began. It ispossible that they followed the trawlsto the seabed, although I have no evi-dence to support this hypothesis.

Bottlenose dolphins are endemic tomany coastal areas throughout theworld’s tropic and temperate waters(Jefferson et al., 1993). They are re-garded as catholic feeders, consuming

60(3), 1998 11

Figure 2.—Diagrammatic representation of prawn trawl, codend, location of camera and lights, and position of dolphins.

a range of fish and cephalopods (Cork-eron et al., 1990; Santos, et al., 1994)and have frequently been observed tocongregate and feed in close proximityto trawling operations (e.g. Leather-wood, 1975; Corkeron et al., 1990; Hilland Wassenburg, 1990; Shane, 1990).This latter observation is a commonoccurrence in NSW and in over 175 and100 days spent onboard prawn and fishtrawlers, respectively (operating both atnight and during the day), I have regu-larly observed bottlenose dolphins at thesurface either (1) removing catch fromcodends while the trawls were retrievedor alternatively, (2) scavenging the catchdiscarded during sorting. Although thereare no published records of the types ofspecies consumed by dolphins duringthese encounters in NSW, in another studyCorkeron et al. (1990) examined the be-

havior of bottlenose dolphins feeding onby-catch discarded from prawn trawlersin Moreton Bay, QLD and concluded thatwhile individual preference played a ma-jor role in determining what was eaten,several species, including whiting, Sillagomaculata, flathead, Platycephalus sp., andsquid, Loligo spp., were often consumedwhile crustaceans were always ignored.

In the present study, repeated ma-nipulation of the codend and the asso-ciated routine feeding pattern was ob-served to occur for up to 20 minutesduring each tow. It was not possible toquantify the amount of small organismsreleased from the codend nor theamount of fish consumed by the dol-phins during this period, however, pre-vious studies discussed by Shane (1990)suggest that an adult bottlenose dolphinmay consume between 4% and 6% of

their body mass daily. A mature dolphin,2.5 m long and weighing approximately200 kg (Mead and Potter, 1990), there-fore, could consume up to 12 kg of fishper day or the equivalent of 300 of thesmall whiting (40 g) or other similar-sized individuals of species that fre-quently appear in by-catches of trawl-ers working throughout the geographicrange of the NSW oceanic prawn-trawlfishery (Broadhurst and Kennelly,1997). This information, when consid-ered along with the size of the prawn-trawl fleet in NSW (300 vessels) andtheir gear configuration (3 nets) illus-trates the potential for dolphins to de-rive a substantial nutritional benefitfrom interactions with prawn trawls.

Despite the evidence to suggest thatdolphins occur in close proximity totrawling operations off NSW and inter-

12 Marine Fisheries Review

Figure 3.—Frames of (A) one dolphin (top right) approaching the posterior section of the codend (bottom), (B)dolphin (top) pushing the posterior section of the codend (bottom) forwards, and (C) two dolphins (side view - topleft) feeding on small fish escaping from the codend. Because visibility was low and the frames were taken fromvideo, picture quality is very poor.

act with the trawls (either at the surfaceor on the seabed), there is little indica-tion that this behavior results in anymortality due to entanglement. For ex-ample, as part of a 3-year observer-based study to examine the by-catch ofprawn trawlers in NSW (Liggins et al.,1996; Liggins and Kennelly, 1996;Kennelly et al., 1998), a total of 579fishing trips were sampled with norecord of any cetacean deaths. In addi-tion, interviews and discussions withcommercial fishermen have revealedthat such deaths are apparently quiterare. These observations are supportedby previous studies examining the in-teraction between bottlenose dolphinsand commercial fishing operations (e.g.Corkeron et al., 1990).

While it is evident that dolphins regu-larly aggregate around trawling opera-tions in NSW, the absence of any datadescribing the feeding behavior andcomposition of diets precludes any es-

timation of the extent or scale of anyinteractions. Nevertheless, assuming thepotential for at least some interaction,the evidence presented here may haveimportant implications for quantifyingthe selectivity of prawn trawls and esti-mating rates of by-catches from prawntrawlers. For example, it is evident thatdolphins interacting with the trawl dur-ing fishing can cause significant changesin the fractional mesh openings in thecodend resulting in alterations to theoverall selectivity of the trawl (espe-cially for prawns). More importantly,however, the scavenging and associatedrelease of by-catch from the codend (ei-ther at the surface or on the seabed) rep-resents one component of fishing-in-duced (F) mortality (see Chopin et al.,1996) that can not be easily estimated,but which nevertheless may contributequite significantly to the overall fish-ing mortality of particular by-caughtspecies (e.g. whiting). The potential for

this type of bias has generally been ig-nored in studies that have quantified by-catches, since it is almost always as-sumed that the catch landed at the endof each tow represents the total catchcaught and retained during fishing.Given the observations made in thepresent study, future research into gearperformance, selectivity and quantifica-tion of by-catches may benefit fromsome assessment of the potential extentof such interactions.

One relatively inexpensive method ofquantifying the extent of the interactionobserved in the present study would beto conduct paired gear comparisons,using twin or tripled rigged trawls(Broadhurst and Kennelly, 1997) witha large-meshed, rigid cover or cage at-tached to one net so that it encompassesand extends beyond the posterior sec-tion of the codend to prevent dolphinsfrom interacting with the main codend.The catch from this “modified” trawl

60(3), 1998 13

14 Marine Fisheries Review

Figure 4.—Diagrammatic representation of proposed modification to posterior sec-tion of codends to prevent interaction of dolphins.

could then be compared to the normalcommercially rigged “control” net. As-suming no differences in the fishingperformances or selectivities of the twonets (due to the presence or absence ofthe cover), which in any case could betested for, the differences in catch ratesacross replicate tows should provide anestimate of the component of by-catchreleased due to interactions by dolphins.

Commercial fishermen in NSW mayalso consider using similar sorts ofmodifications to the posterior sectionsof their codends (behind the compositesquare-mesh panel) on a regular basisto prevent dolphins from manipulatingthe codend and inadvertently releasingcommercially important prawns andcephalopods. Figure 4 shows one pos-sible modification, comprised of a panelof large, heavy mesh (e.g. 4 mmbraided, mesh size 90 mm) sewn aroundthe posterior section and extending forup to 1 m past the end of the codend,that may prevent physical contact bydolphins. Variations of this type ofmodification have been successfullyused by fishermen for many years toprevent sharks from damaging thecodend and catch during fishing, withno recorded deaths of cetaceans due toentanglement. Alternatively, it may befeasible to examine the utility of trawl-mounted, battery-operated acoustic de-terrents, similar to those used to reducethe by-catch of dolphins in gillnet fish-eries in the northeastern United States(Lien, 1995). Further research on a fish-ery-specific basis is needed, however,to assess the extent to which dolphinsrely on food from trawls during fishingand, if any modifications to trawls toprevent removal of by-catch wouldnegatively influence their abundances.

Acknowledgments

This work was funded by the Aus-tralian Fishing Industry Research andDevelopment Corporation (Grant 93/180). Thanks are extended to GaryAnderson for the use of his vesselValerie Dawn, John Matthews for tech-nical assistance, Sandie Ingleby forhelpful comments on the video footage,and Nick Otway and Steve Kennelly forcritically reviewing the manuscript.

Literature CitedAndrew, N. L., and J. G. Pepperell. 1992. The

by-catch of shrimp trawl fisheries. Oceanogr.Mar. Biol. Annu. Rev. 30:527–565.

Blaber, S. J. M., D. A Milton, G. C. Smith, andM. J. Farmer. 1995. Trawl discards in the di-ets of tropical seabirds of the northern GreatBarrier Reef, Australia. Mar. Ecol. Prog. Ser.127:1–13.

________ and T. J Wassenburg. 1989. Feedingecology of the piscivorous birds Phala-crocorax varius, P. melanoleucos and Sternabergii in Moreton bay, Australia: diets anddependence on trawler discards. Mar. Biol.(Berl.) 101:1–10.

Broadhurst, M. K., and S. J. Kennelly. 1996. Ef-fects of the circumference of codends and anew design of square-mesh panel in reducingunwanted by-catch in the New South Walesoceanic prawn-trawl fishery, Australia. Fish.Res. (Amst.) 27:203–214.

__________ and S. J. Kennelly. 1997. The com-posite square-mesh panel: a modification forreducing unwanted by-catch and increasingcatches of prawns throughout the New SouthWales oceanic prawn-trawl fishery. Fish. Bull.95:653–664.

Chopin, F., Y. Inoue, and T. Arimoto. 1996. De-velopment of a catch mortality model. Fish.Res. (Amst.) 25:377–382.

Corkeron, P. J., M. M. Bryden, and K. E.Hedstrom. 1990. Feeding by bottlenose dol-phins in association with trawling operations inMoreton Bay, Australia. In S. Leatherwood andR. R. Reeves (Editors), The bottlenose dolphin,p. 329–336. Acad. Press, San Diego, Calif.

Hill, B. J., and T. J. Wassenburg. 1990. Fate ofdiscards from prawn trawlers in Torres Strait.Aust. J. Mar. Freshwater Res. 41:53–64.

Jefferson, T. A., S. Leatherwood, and M. A.Webber. 1993. Marine mammals of the world.FAO Species Identification Guide. UNEnviron. Programme, Rome, Italy, 320 p.

Kennelly, S. J. 1995. The issue of bycatch inAustralia’s demersal trawl fisheries. Rev. FishBiol. Fish. 5:213–234.

________ , G. W. Liggins, and M. K. Broadhurst.1998. Retained and discarded by-catch from

oceanic prawn trawling in New South Wales,Australia. Fish. Res. (Amst.) 36:217–236.

Leatherwood, S. 1975. Some observations offeeding behavior of bottle-nosed dolphins(Tursiops truncatus) in the northern Gulf ofMexico and (Tursiops cf T. gilli) off southernCalifornia, Baja California, and Nayarit,Mexico. Mar. Fish. Rev. 37(9):10–16.

Lien, J. 1995. Conservation aspects of fishinggear: Cetaceans and Gillnets. In Solving by-catch—considerations for today and tomor-row, p. 219–224. Alaska Sea Grant CollegeProgram Report No. 96-03, Univ. AlaskaFairbanks.

Liggins, G. W., S. J. Kennelly, and M. K.Broadhurst. 1996. Observer-based survey ofby-catch from prawn trawling in Botany Bayand Port Jackson, New South Wales. Mar.Freshwater Res. 47:877–888.

________ and S. J. Kennelly. 1996. By-catchfrom prawn-trawling in the Clarence Riverestuary, New South Wales, Australia. Fish.Res. (Amst.) 25:347–367.

Mead, J. G., and C. W. Potter. 1990. Natural his-tory of bottlenose dolphins along the centralAtlantic coast of the United States. In S.Leatherwood and R. R. Reeves (Editors), Thebottlenose dolphin, p. 165–195. Acad. Press,San Diego, Calif.

Santos, M. B., G. J. Pierce, H. M. Ross, R. J.Reid, and B. Wilson. 1994. Diets of small ce-taceans from the Scottish coast. ICES C.M./N:11, 16 p.

Shane, S. S. 1990. Comparison of bottlenosedolphin behaviour in Texas and Florida, witha critique of methods for studying dolphinbehavior. In S. Leatherwood and R. R. Reeves(Editors), The bottlenose dolphin, p. 541–558.Academic Press, San Diego, Calif.

Wassenburg, T. J., and B. J. Hill. 1987. Feedingby the sand crab Portunus pelagius on mate-rial discarded from prawn-trawlers in More-ton Bay, Australia. Mar. Biol. (Berl.) 95:387–393.

________ and ________ . 1990. Partitioning ofmaterial discarded from prawn trawlers inMoreton Bay. Aust. J. Mar. Freshwater Res.41:27–36.