SICB 2008

Correlations of Bite Performance with Head and Carapace Morphometrics in Green Sea Turtles (Chelonia mydas)

Marshall, C.D.
Department of Marine Biology, Texas A&M University at Galveston, marshalc@tamug.edu

Adult green sea turtles (Chelonia mydas) are unusual relative to other sea turtles in that they are predominately herbivorous.  This herbivorous diet is reflected in the serrated morphology of their rhamphotheci, and likely in the relative morphometrics of their heads. Recent bite performance data in loggerhead turtles (Chelonia mydas), which are known for their durophagous capability, have demonstrated that bite force is correlated with head width. Therefore, the objective of this study was to characterize head and carapace morphometrics, and correlate these data with bite force performance in green sea turtles.  It was predicted that maximum bite force in green sea turtles would be less in magnitude relative to loggerheads, but would also be positively correlated with head width. Therefore, carapace & head morphometrics, and bite force performance were collected from free-ranging green sea turtles (N=63) from Punta Abreojos, Baja California Sur, Mexico. Subjects ranged from 13.6 to 149.7 kg, with a mean straight carapace length and width of 58.9 cm (S.D. ±12.5) and 46.4 ±8.0, respectively.  A bite force apparatus that incorporated a piezoelectric force transducer, with bite plates customized for sea turtles, was used to collect bite force data from subjects.  Bite force was measured at the rostral most tips of the jaws. The mean maximum bite force was 152.3±64 N and ranged from 36 to 313 N.  Mean head width, head height, and head length were 89.2±15.4, 56.1±29.2, and 118.4 ±20.5 mm, respectively.  Bite force was lower in magnitude than reported for loggerhead turtles, as predicted.  A stepwise multi-linear regression demonstrated that head width was also the best predictor of bite force (P < 0.01; R2 = 0.32).
 

SICB 2008

Comparative Feeding Performance in Odontocetes

Kane, E.; Marshall, C.D.
Department of Marine Biology, Texas A & M University at Galveston, 5007
Avenue U, Galveston, TX 77551, USA
ekane@tamu.edu

Suction is the ancestral feeding mode among aquatic vertebrates.  However, the terrestrial origin of odontocetes resulted in an ancestral ram feeding morphology.  The recent evolution of suction feeding in some species has given rise to morphologies such as blunt and wide rostra, reduced dentition, and a circular oral aperture, which converge with other known aquatic suction feeders. Unfortunately, few data exist that quantify feeding performance in odontocetes.  Therefore, the objective of this study was to directly measure intraoral suction generation during controlled feeding trials in presumed suction and ram feeding odontocetes (beluga whales, Delphinapterus leucas, and Pacific white-sided dolphins, Lagenorhynchus obliquidens). Analysis of 883 feeding events from 6 belugas and 7 Pacific white-sided dolphins demonstrated that the maximum negative suction pressure measured in belugas in this study was ­122 kPa, which was 4.5 times greater than the maximum value generated by Pacific white-sided dolphins (­27 kPa).  Hydraulic jetting was also measured in belugas with approximately equal magnitude (128 kPa).  Although the use of suction in belugas is well known, these are the first data to measure the magnitude; the suction capability of belugas appear to be comparable to walruses. Despite long rostra, numerous teeth, and a limited ability to occlude lateral gape, Pacific white-sided dolphins are also capable of using suction, but to a lesser extent than belugas. Their suction capability is likely limited to the intraoral transport of prey whereas in belugas it is a primary method of prey capture.
 

SICB 2008

Ontogeny of Bite Performance in Loggerhead Sea Turtles (Caretta caretta)

Alejandra Guzmán & Christopher D. Marshall

Although it is perceived that loggerheads are powerful biters, no data exists that characterizes their bite performance.  Since loggerheads are durophagous, an understanding of their bite performance would provide insight on feeding behavior, trophic ecology, as well as an understanding for interactions with longline fishing gear. Loggerhead bite force was measured in four known age classes of captive-reared juveniles, and in larger wild and captive turtles of unknown age classes.  Mean maximum bite force of hatchlings less than six months was 3.1 N with monthly mean increases of 0.5 N between the ages of 3 and 5 months. Mass was the best predictor of hatchling bite force (Adj. R2 = 0.612).  Mean maximum bite force of 9, 20, 34, and 36 month old loggerheads were 22 N, 129 N, 358 N, and 457 N, respectively.  Hatchling bite force was less than 1% of bite force attained by turtles three years of age.  Bite force in larger sub-adult and adult turtles (SCL 59-99 cm) ranged from 382 to 1,221 N.  In all age classes, maximum bite force had a positive linear relationship with all head and body morphometrics (P<0.001).  Head width was the best predictor of bite force throughout juvenile ontogeny (Adj. R2 = 0.936). Morphometric predictions of bite force in large, wild and captive loggerheads were higher than observed values in all cases.  A power function curve estimation provided the best fit of bite force data throughout ontogeny (Adj. R2 = 0.912).
 

SICB 2008

High Speed Video Analyses of Loggerheads Interacting With Baited Longline Hooks

Alejandra Guzmán & Christopher D. Marshall

Interactions between marine turtles and the longline fishing industry are a growing concern. Feeding biomechanics (i.e., bite performance and feeding kinematics), foraging behavior, and diet preferences are likely candidates for loggerheads vulnerability to capture in the longline fishery. A motion analysis study was conducted to investigate how loggerheads interacted with five barb-less, squid-baited longline hooks (J with 25š offset, 16 circle no offset, 16 circle with 10š offset, 18 circle no offset, and 18 circle with 10š offset). Frontal and lateral perspectives for each hook treatment were videotaped and analyzed for all loggerheads (SCL 40-55 cm; n=7, 50 feeding trials per individual). Nine kinematic variables (e.g, maximum gape distance) and three behavioral variables (e.g, frequency of strikes and hooking location) were assessed in relation to hook treatment.  These data were used to test the hypothesis that loggerhead feeding behavior is modulated relative to hook type, hook size and hook offset angle. The mean loggerhead kinematic profile was 6.45 cm (+0.91) for maximum gape, 0.73 s (+0.37) for duration of gape opening, 0.13 s (+0.06) for duration of gape closing, 0.86 s (+0.38) for total feeding cycle duration , 4.27 cm (+0.93) for maximum hyoid depression, and 0.86 s (+0.38) for time to maximum hyoid depression.  No difference was detected between hook treatments in all kinematic variables analyzed from the frontal perspective (p>0.05). Frequency of strikes (mean 1.03+0.19) did not vary between hook treatments or hooked and non-hooked interactions. The hooking percentage was lowest in the 16 circle straight (29.6%) and 18 circle with 10š offset (29.8%) hooks. The highest hooking percentage (45%) occurred when turtles interacted with 16 circle hooks with 10š offset. The smallest difference in hooking percentage was observed in 18 circle no offset and 18 circle with 10š offset hooks. The 16 circle hook with 10š offset had the highest percentage of hooking in the throat (48.3%) while the 18 circle no offset hook resulted in the lowest percentage of hooking in the throat (21.1%). These investigations of simulated loggerhead-longline fishery interactions will provide data that can be used to develop or modify longline fishery gear that may potentially reduce loggerhead bycatch.
 

International Sea Turtle Symposium, Loreto, Baja California Sur, Mexico 2008

Effects of Hook Type on Kinematic and Behavioral Variables of Loggerheads (Caretta caretta) Interacting with Baited Longline Hooks

Alejandra Guzmán & Christopher D. Marshall
asalazarguzman@gmail.com

Interactions between marine turtles and the longline fishing industry are a growing concern. Feeding biomechanics, foraging behavior, and diet preferences are likely candidates for loggerheads¹ vulnerability to capture in the longline fishery. A kinematic study on 7 loggerheads (SCL 40-55 cm) was conducted to investigate how these turtles interacted with five barb-less, squid-baited longline hooks (J with 25šoffset, 16 circle no offset, 16 circle with 10šoffset, 18 circle no offset, and 18 circle with 10šoffset).  Nine kinematic variables and five behavioral variables were assessed in relation to hook treatment.  These data were used to test the hypothesis that loggerhead feeding behavior is modulated relative to hook type, hook size and hook offset.  No difference was detected between hook treatments in all kinematic variables analyzed (P>0.05) indicating loggerhead feeding behavior may be stereotypical.  Maximum gape, maximum hyoid depression, gape angle, time to maximum gape, time to maximum hyoid, gape closing time, total feeding time, open gape angle velocity and closing gape angle velocity were 6.5 + 0.9 cm, 4.3 + 0.9 cm, 48.4š + 7.9š, 0.7 + 0.3 s, 0.9 +, 0.4 s, 0.1 + 0.06 s, 0.9 + 0.4 s, 611 + 182 deg/s, and -854 + -239 deg/s respectively.  Mean frequency of strikes did not vary between hook treatments or hooked and non-hooked interactions. Only 33% of all interactions resulted in hooking incidents.  The proportion of hooking compared to non-hooking incidents was significantly different in these hook treatments (P<0.001). The overall hooking percentage was lowest in the 16 circle hook with no offset (29.7%) and 18 circle hooks with 10šoffset (29.9%), while the highest percentage of hooking occurred in the 16 circle hook with 10šoffset (45.2%).  A more specific inspection of hooking location (mouth or throat) demonstrated that the proportion of turtles hooked in the mouth relative to turtles hooked in the throat in all hook treatments (P=0.001).  The 16 circle hook with 10š offset had the highest percentage of throat hooking (48.3%); the 18 circle no offset hook resulted in the lowest percentage of throat hooking (21.3%).  No difference was detected in head orientation towards the hook (right, straight, or left) in 16 and 18 circle hooks with no offset (P>0.05).  There was a significant difference in the proportion of turtles that oriented themselves towards and away from the J hook with 25š offset and the 16 and 18 circle hooks with 10š offset (P<0.001).  When interacting with J hooks with a 25š offset, turtles mostly oriented themselves away from the hook offset (e.g, hook barb) (77.2%); however, when interacting with the 16 and 18 circle hooks with 10š offset, most turtles oriented themselves towards the hook offset (56.9% and 50.7%). These investigations of simulated loggerhead-longline fishery interactions may provide data that can be used to develop or modify longline fishery gear that may potentially reduce loggerhead bycatch.
 

SICB Abstract 2007

Functional Morphology of Inking Behavior in Pygmy & Dwarf Sperm Whales

Marshall, CD and Moss AL
Texas A&M University at Galveston, marshalc@tamug.edu

Pygmy & dwarf sperm whales (Kogia breviceps and K. sima) demonstrate an unusual inking behavior similar to cephalopods. Such behavior is unknown among other cetaceans. The source of ink is hypothesized to be the production of a large volume of watery stool (12-50 l). Evidence suggests that the mechanism for ink production kogiids is a derived digestive physiology in the enlarged, sac-like colon. A systematic investigation of kogiid gastrointestinal (GI) tracts has not been conducted. Therefore, this study investigated the morphology and cytology of the kogiid GI tract. Five GI tracts were collected through the SEUS Stranding Network. Standard morphometrics of the GI tract were taken (length, diameter, and volume), and samples were process for histology, and then stained with a modified Masson’s trichrome. Morphometrics of epithelia were taken and enterocytes were identified and quantified (e.g., height & width of villi, thickness of mucosa & adjacent tissue layers, and the factor of the increase in absorptive surface area). The mean length of the jejunum and colon were 23.3 m and 2.26 m, respective. External lengths of the duodenum and ileum were arbitrary due to the lack of external distinctions between adjacent segments. Mean diameter of the duodenum, jejunum, and colon, were 2.28, 1.93, and 7.84 cm, respectively. Mean volume of the duodenum, jejunum, and colon were 220, 8,693.2, & 1,398.9 ml³, respectively. Maximum colon diameter and volume was 16.5 cm & 1771.5 ml³. Plicae and villi of the jejunum and colon were remarkably limited, short and exhibited a low surface area relative to bottlenose dolphins. Mucous cells dominated both the jejunum and the colon. We hypothesize that the normal function of the colon, has been disrupted to produce large volumes of ink.
 

SICB Abstract 2007

Correlations of Bite Performance with Head Morphometrics in Juvenile Loggerhead turtles (Caretta caretta): Preliminary Data

Salazar, AX, Salazar@neo.tamu.edu
Marshall, CD, Texas A&M University at Galveston

Incidental bycatch of loggerhead sea turtles from the pelagic longline fishery is a serious concern due to recent data that demonstrates a greater than predicted turtle bycatch globally. Little is known about loggerhead bite performance, but the propensity of loggerheads to bite forcefully may make them vulnerable to becoming hooked in the mouth or throat. Bite force and associated morphometric measurements were taken from juvenile captive-reared loggerheads at the NOAA turtle facility in Galveston, TX. In addition to bite force measurements, straight carapace length (SCL), straight carapace width (SCW), head width (HW), head height (HH), head length (HL) and body-mass were also recorded from individual of the 2004 (n=180) and 2005 (n=29) age classes. The mean maximum bite force was 140 N for the 2004 age class and 24 N for the 2005 age class. Mean morphometric values for the 2004 age class were 30.7 cm SCL, 25.5 cm SCW, 62.3 mm HW, 53.9 mm HH, 93.5 mm HL and mass 4032 grams. Mean values for the 2005 age class were 12.0 cm SCL, 9.7 cm SCW, 28.3 mm HW, 24.7 mm HH, 45.0 mm HL and mass was 266 grams. Maximum and mean bite force measurements are positively correlated to all body and head morphometrics (p<0.001). Head width is the best predictor of maximum (adjusted R2= 0.899; maximum bite force=3.481(HW) – 76.497) and mean bite force (adjusted R2= 0.893; mean bite force=3.215(HW) – 71.158) in these juvenile loggerheads. Investigations of loggerhead bite performance may provide data that can be used to modify longline fishing gear and diminish sea turtle bycatch.

SICB Abstract 2007

Feeding Performance of Juvenile Red Snapper (Lutjanus campechanus) From Two Habitats

Case, J.E. Texas A&M University at Galveston, casejanelle@neo.tamu.edu
Marshall, C.D.

Juvenile red snapper are attracted to structure and settle onto low profile reefs, which serve as nursery grounds. Little is known about their biology during this time. Recent studies from a shell bank in the NW Gulf of Mexico have shown higher growth rates for fish located on adjacent mud habitats, perhaps due to varied prey availability and abundance. For this study juveniles were collected from a shell ridge (on ridge) and adjacent mud areas (off ridge) on Freeport Rocks, TX. Feeding morphology and kinematics were characterized and compared between the two habitats. A dynamic lever model was used to make predictions about feeding mechanics, and was validated using kinematic profiles obtained from high-speed videos of prey capture events. The model reasonably predicted kinematic profiles of jaw closure for maximum gape velocity and displacement. Model output suggested that on ridge fish have faster jaw closing velocities than off ridge fish. Off ridge fish therefore have greater mechanical advantage and longer closing durations. Analysis of prey capture events indicated that on ridge fish demonstrate greater velocities and larger displacements of the jaws than off ridge fish. Shape analysis was used to further investigate habitat effects on morphology. Off ridge fish differed from on ridge in possessing a deeper head and body. Results from model simulations, kinematic profiles, personal observations, and shape analysis all compliment the conclusion that on ridge fish exhibit suction feeding behavior, whereas fish from off ridge use biting behavior. These results suggest that habitat disparity, perhaps available prey composition, has generated variations in feeding mechanics and consequently feeding behavior. This may ultimately affect growth rate and indicate nursery quality.
 

SICB Abstract 2006

Using Morphology to Predict Feeding Behavior: A Preliminary Study of Juvenile Red Snapper (Lutjanus campechanus) Feeding

Case, J.E. Texas A&M University at Galveston, casejanelle@neo.tamu.edu
Marshall, C.D. Texas A&M University at Galveston

Red snapper are an important commercial species in the Gulf of Mexico. However, their biology and feeding mechanics are virtually unknown. The goals of this study were to, (1) characterize the feeding biomechanics of juvenile red snapper and, (2) to test the hypothesis that the jaw biomechanics differ between subjects from different size classes and habitats. Juvenile red snapper from three size classes (<49 mm, 50-69 mm, and >70 mm SL) and two habitats (on-ridge & off-ridge) were collected from the northwestern Gulf of Mexico off the coast of Texas. Fourteen morphometric measurements of the lower jaw and main closing adductor muscles, A2 and A3, were recorded. These data were used in the biomechanical model MandibLever 3.0 (M.W. Westneat), which provided dynamic outputs of predicted jaw function, such as effective mechanical advantage (EMA) and jaw velocity (VR). Outputs from the model were used to predict and characterize the feeding biomechanics from the three size classes and two habitats. Overall, the model reported a low EMA and high VR, suggesting that juvenile red snapper are primarily suction feeders. This suggestion is supported by kinematic feeding trials of live juvenile red snapper. No significant differences among size classes (p > 0.05) or between habitats (p > 0.05) were observed in EMA or VR. However, as body size increased, VR tended to decrease, and EMA tended to increase. Although our data is for juveniles only, the trends in EMA and VR suggest a shift from primarily suction feeding as juveniles to primarily biting as adults. This trend is supported by kinematic feeding trials of juvenile red snapper, modeling data from a few adult red snapper, as well as observations of feeding behavior in adults.
 

SICB Abstract 2005

Behavioral Performance of Feeding in Kogia and Tursiops (Odontoceti: Cetacea):
Ram vs. Suction Feeding Strategies

BLOODWORTH, B.E. Texas A&M Univ., Galveston, bloodwob@neo.tamu.edu
MARSHALL, C.D. Texas A&M Univ., Galveston

The feeding performance of two bottlenose dolphins (Tursiops truncatus) and one juvenile pygmy and dwarf sperm whale (Kogia breviceps and K. sima) was compared during controlled feeding trials. Feeding behavior was characterized and a Ram-Suction Index (RSI) measured. The feeding behavior of both genera was composed of four phases: preparatory (Phase I), jaw opening (Phase II), gular depression (Phase III) and jaw closing (Phase IV). The mean Kogia total feeding cycle duration (533 ms; SD±150) was shorter than the mean for Tursiops (847 ms; SD±349). The mean maximum gape angle was greater for Kogia (56.6º; SD±3.3) than for Tursiops (25.5º; SD±8.5). The mean Kogia RSI (-0.04; SD±0.378) was significantly less (p<0.001) than the mean Tursiops RSI (0.71; SD±0.14), indicating a suction and ram-based strategy for Kogia and Tursiops, respectively. The significantly shorter (p<0.05) Kogia jaw opening phase (172 ms; SD±75) supports the RSI results. Tursiops displayed two feeding behaviors, a Type I in which the mandibles open near to a food item and a Type II in which gape was near maximum while the subject was greater than one meter away from food. Total feeding cycle duration and RSI of Type I behavior (693 ms; SD±273 and 0.65; SD±0.15) were significantly different (p<0.001) from Type II behavior (1135 ms; SD±297 and 0.79; SD±0.05). Significant differences in RSI also existed between Kogia (-0.04; SD±0.378), Type I (0.65; SD±0.15) and Type II (0.79; SD±0.05) feeding behaviors. The results suggest odontocetes demonstrate a wide range of feeding strategies along the ram-suction continuum not previously characterized. Variability among individuals can be high and multiple feeding behaviors are possible even within a single genus (i.e., Tursiops).

16th Biennial Conference on the Biology of Marine Mammals Abstract -- 2005:

Behavioral Performance of Suction Feeding in Harbor Seals (Phoca vitulina)

Marshall, Christopher. D. Texas A&M University at Galveston, Department of Marine Biology
Dehnhardt, Guido Ruhr-Universität Bochum, Lehrstuhl Allgemeine Zoologie & Neurobiologie

As part of a comparative study on the feeding performance of pinnipeds, suction generation capability was measured in harbor seals.  Depending upon their geographic location suction is thought to be an important component of the feeding repertoire in harbor seals.  However, behavioral performance studies that investigate suction generation have not been conducted.  Harbor seals (N=2) were presented pieces of fish in a feeding apparatus in which they had a choice to ingest food projecting from a hole drilled through a sheet of plexiglass by biting, or from a recessed cylinder using suction.  A Millar MPC 350 pressure transducer was inserted into one of the cylinders of the feeding apparatus and pressure data was collected with a portable electrophysiological recording system.  Feeding behavior was videotaped and synchronized with physiological data using a pair of flashing LED lights whose optical pattern corresponded to a generated square wave pattern recorded as a separate channel.  Harbor seals used suction, biting, and some hydraulic jetting when feeding from the apparatus.  The mean suction force recorded was -10.0 kPa (S.D. + 6.54).  The maximum suction pressure measured over 137 feeding events was 31.3 kPa.  The mean duration of suction events were 0.5 s (S.D. + 0.36) and ranged from 0.13 s to 2.06 s.  Suction events were occasionally preceded by a small increase in pressure (preparatory phase) followed by relatively large negative pressures (suction).  In some cases a small increase in pressure was also noted after a suction event.  These supplemental phases to the main suction event presumably served to remove excess water from the mouth either just before or after a suction event.  Suction capability of harbor seal were considerably less compared to suction specialists such as bearded seals and walruses, but were strong enough to be an important part of the overall feeding repertoire. 
 

SICB Abstract 2005:

Suction and Hydraulic Jetting Forces Produced by Feeding Bearded Seals (Erignathus barbatus)

Marshall, C. D. Texas A&M University at Galveston, USA; marshalc@tamug.edu
Kovacs, K. M. Norwegian Polar Institute, Tromsø, Norway
Lydersen, C., Norwegian Polar Institute, Tromsø, Norway

Bearded seals are thought to use suction as their primary feeding mode. This suggestion was, until recently, based only on anecdotal data. However, recent studies have confirmed that both suction and hydraulic jetting are important feeding behaviors of this species; the magnitude of these forces remained unknown. This study characterized suction and hydraulic jetting behaviors of bearded seals, and measured their magnitudes using a Millar MPC 350 pressure transducer inserted into a feeding apparatus. Pressure data from 70 feeding trials were collected with a portable electrophysiological recording system. Feeding behavior was videotaped and synchronized with physiological data using a pair of flashing LED lights whose optical pattern corresponded to a generated square wave pattern recorded as a separate channel. Bearded seals (N=2) used both suction and hydraulic jetting when feeding from the apparatus. The mean suction force recorded was -47.1 kPa (S.D.+ 25.7) and ranged from -3.2 to -108 kPa (N=392 suction events). The mean hydraulic jetting force recorded was 8.9 (S.D.+ 5.45) and ranged from 3.2 to 37 kPa (N=63 hydraulic jetting events). The mean duration of suction events (0.29 s, S.D.+ 0.12) was significantly greater (p <0.001) than the mean duration of hydraulic jetting events (0.13 s, S.D.+ 0.07). Suction events were often preceded by a small increase in pressure (preparatory phase) followed by relatively large negative pressures (suction phase). Suction capability of bearded seals are comparable to walruses, and validate anecdotal data that bearded seals are suction foragers. However, hydraulic jetting is an important component of the feeding repertoire of bearded seals.
 

Abstract from a Presentation to the Society of Marine Mammals in 2002:

Morphology of the Bearded Seal (Erignathus barbatus) Muscular-Vibrissal Complex

Christopher D. Marshall, University of Washington, Box 357446, Seattle, WA 98195, USA
Kit Kovacs, Norwegian Polar Institute, 9296 Tromsø, Norway
Christian Lydersen, Norwegian Polar Institute, 9296 Tromsø, Norway

Bearded seals possess an elaborate array of mystacial vibrissae that are assumed to be involved in the identification and manipulation of benthic prey. However, little information is available regarding the anatomy of the vibrissae, facial myology or the innervation of these structures – so there is little evidence available to support or dispute this hypothesis. This study reports on the facial myology, gross innervation, and vibrissal histology of the bearded seal. Specimens were obtained from legal hunts in Barrow, Alaska and Norway. The muzzle is comprised primarily of three enlarged muscular layers, the levator nasolabialis (superficial), maxillonasolabialis (intermediate), and lateralis nasi (deep). These muscles possess bony origins but soft tissue insertions that presumably give the muzzle increased mobility. A complex anastomosing network of CN VII innervates these facial muscles. The disproportionately large infraorbital nerve of CN V courses toward the muzzle and divides into numerous branches that penetrate the external capsule of every vibrissa. Each vibrissal follicle spans the depth of the dermis. Each follicle possesses a blood filled ring sinus, located between the dense external capsule and the inner wall, that encircles the hair shaft. A trabecular cavernous sinus is located apically and basally to the ring sinus. The inner wall is comprised of a ringwulst, mesenchymal sheath, glassy membrane, and outer root sheath. The inner wall joins the external capsule apically to form a pore through which the hair shaft emerges. Nerve fibers penetrating the external capsule course through the basal trabeculae to terminate, presumably as mechanoreceptors, on the inner wall adjacent to the ring sinus. Thus, the anatomical evidence suggests that the vibrissae of bearded seals are highly developed tactile organs that can be mobilized and serve to help locate, identify, and access benthic prey.