Differences among life-history stages, morphology, and prey availability influence sea turtle feeding strategies as well as their geographical and temporal distributions. Adult feeding preferences vary by species, with Kemp's ridley (Lepidochelys kempii) and loggerhead (Caretta caretta) sea turtles considered opportunistic benthic carnivores. In contrast, adult hawksbill (Eretmochelys imbricata) and leatherback (Dermochelys coriacea) sea turtles are typically specialists on sponges and gelatinous prey, respectively, whereas the green sea turtle (Chelonia mydas) is primarily herbivorous (reviewed by Bjorndal 1997; Jones and Seminoff 2013).

All life stages of the critically endangered (International Union for Conservation of Nature [IUCN] 2015) Kemp's ridley sea turtle seasonally inhabit the northwestern Gulf of Mexico, but the area is utilized primarily as developmental and foraging habitat during spring and summer months (Landry and Costa 1999; Landry et al. 2005; Seney and Landry 2011). The region is also used as adult migratory and internesting habitat (Seney and Landry 2008). An ontogenetic shift is believed to occur at about 20–25 cm straight carapace length (SCL), or about 2 yrs of age, when juveniles leave pelagic Sargassum mats for coastal areas and begin to forage on benthic invertebrates (Schmid and Witzell 1997; Landry and Costa 1999; Snover et al. 2007), although there may be some plasticity in this shift (e.g., Putman and Mansfield 2015). An additional ontogenetic shift in habitat use and prey may occur at approximately 40–50 cm SCL, as evidenced by an apparent growth spurt at this size and differences in size class frequencies between some habitats (summarized in Schmid and Barichivich 2006). Maturity is presumably reached at approximately 58–60 cm SCL (Schmid and Barichivich 2006).

Geographical and size-based differences are apparent in the diets of post-pelagic Kemp's ridleys foraging in US waters. Small juveniles from New York concentrate foraging efforts on walking crabs, particularly the portly spider crab (Libinia emarginata) and Atlantic rock crab (Cancer irroratus; Burke et al. 1993, 1994). Larger conspecifics sampled in Virginia (Seney and Musick 2005), Georgia (Frick and Mason 1998), and Texas (Shaver 1991; Werner 1994) consumed a large amount of common blue crab (Callinectes sapidus) and other portunid (swimming) crabs, in addition to walking crabs such as spider crabs (Libinia spp.), mottled purse crab (Persephona mediterranea), and stone crabs (Menippe spp.). In contrast, immature Kemp's ridleys captured in southwestern Florida rarely consumed portunid crabs and instead have fed primarily on benthic tunicates (Molgula occidentalis) and walking crabs, as well as tube-building polychaete worms (Witzell and Schmid 2005; Schmid and Barichivich 2006). Various molluscs, finfish, shrimp, jellyfish, Atlantic horseshoe crab (Limulus polyphemus), diamondback terrapin (Malaclemys terrapin), and vegetation have also been reported in the species' diet in one or more US location (Shaver 1991; Werner 1994; Frick and Mason 1998; Seney and Musick 2005; Witzell and Schmid 2005). Adult Kemp's ridleys are found almost exclusively in the Gulf of Mexico, where their diet is believed to be similar to that of immature individuals and largely composed of crabs (Shaver 1991; Jones and Seminoff 2013).

Once on the brink of extinction, the Kemp's ridley population has rebounded due in large part to conservation efforts by Mexico and the United States. However, the species' primary Tamaulipas, Mexico, and secondary Texas, United States, rookeries have both exhibited a shift from exponential growth to a flat or perhaps declining trend that began in 2010 (National Marine Fisheries Service [NMFS] and US Fish and Wildlife Service [US FWS] 2015). Hypotheses for this abrupt shift include both a reduced adult female population size due to mortality and increased nesting remigration intervals due to reduced prey availability (Gallaway and Gazy 2014).

By providing insights to a species' ecosystem role, life-history stage, ontogenetic shifts, and interactions with anthropogenic activities (e.g., Shoop and Ruckdeschel 1982; Shaver 1991; Burke et al. 1993; Seney and Musick 2005; Schmid and Barichivich 2006), diet characterization generates a better understanding of sea turtle ecology and can thereby improve management schemes. Recognizing these benefits, a dietary analysis was conducted for Kemp's ridleys caught and entangled in recreational hook-and-line fishing gear along the upper Texas coast to characterize diet of these incidental captures and examine their feeding strategies. Hook-and-line captures have been documented in Texas for several decades (summarized in Seney 2008) and are of increasing concern as a threat to Kemp's ridleys in the northern Gulf of Mexico (Coleman et al. 2013).

METHODS

The National Oceanic and Atmospheric Administration, National Marine Fisheries Service's Sea Turtle Facility (NOAA STF) in Galveston, Texas, United States, has responded to reports of sea turtles caught on recreational hook-and-line along the upper Texas coast, with varying intensity, since 1980. During 2005–2008, Galveston County fishing-pier operators were contacted annually by the NOAA STF and Texas A&M University at Galveston to promote reporting of all sea turtles caught on recreational hook-and-line fishing gear; and informational posters were placed at piers, bait shops, parks, refuges, and other appropriate venues along the upper Texas coast. Reported turtles were rescued as quickly as possible and transported to the NOAA STF for holding, removal of easily accessible hooks, photographs, and collection of morphometric data and fecal samples. Straight carapace length was measured from the nuchal notch to the end of the longest pygal scute (notch-to-tip SCL). Each turtle was transported to The Houston Zoo, Inc Veterinary Clinic, where it was examined, radiographed to detect and evaluate the need to remove ingested hooks, and have any necessary medical procedures performed. Upon their return to the NOAA STF, turtles were held for treatment, as prescribed, and ultimately tagged and released in eastern Jefferson County, Texas, near the Texas–Louisiana border, or 32 km (20 miles) offshore of Galveston Island.

Each hook-and-line–caught and –entangled Kemp's ridley sea turtle was housed individually at the NOAA STF in an appropriately sized tank (Higgins 2003) that was checked for feces prior to daily cleaning. One fecal sample was collected within 0–72 hrs of capture (Burke et al. 1994), washed in a 0.5-mm (#35) brass sieve, and remaining contents preserved in 10% freshwater formalin (Seney and Musick 2005). Food items were subsequently identified to the lowest possible taxonomic level.

Percent frequency of occurrence (%F) was calculated for each prey type and for general prey groups: crustaceans, large molluscs, annelids (polychatete worm tubes), fish, and plants, as well as shell hash and natural debris (including mollusc and common sand dollar, Echinarachnius parma, fragments, molluscs < 1 cm, and rocks) and anthropogenic debris. Further metrics were not calculated because a fecal sample does not constitute a standardized sample unit as compared with a whole digestive tract (Burke et al. 1994; Seney and Musick 2005). Additionally, it should be noted that analysis of feces may result in some bias because undigested, excreted material is not likely to include gelatinous prey or other soft-bodied animals. Statistical analyses were not conducted because of low sample sizes among years and size classes.

RESULTS

Fecal samples were collected from 21 immature Kemp's ridleys following incidental capture on recreational hook-and-line gear (n = 20) or entanglement in monofilament and fishing weights (n = 1) during spring and summer (April–August) of 2005–2008. Seventeen turtles originated from Galveston County fishing piers (16 captures, 1 entanglement), 3 from surf fishermen, and 1 from a boat fishing near a jetty (Fig. 1). Over half (57%) were encountered during April–May, and a third during July–August. Sampled Kemp's ridleys measured 26.1–44.5 cm SCL (notch-to-tip, mean = 33.9 cm, SD = 4.6 cm) and weighed 2.6–12.4 kg (mean = 5.8 kg, SD = 2.4 kg) at the time of capture or entanglement. Ten hook-and-line caught turtles swallowed hooks at the time of capture, whereas the other 10 were hooked in the mouth or jaw. Baits included shrimp and squid, and no artificial lures were documented; however, not all hooks and baits were examined because of removal by anglers and subsequent transfer of the turtle to pier staff or law enforcement before the arrival of a biologist. Apart from capture-related injuries, the only fresh wounds documented were shark bites on 2 individuals (photos examined by P. Motta, pers. comm. to E.E.S.) and what appeared to be a boat-propeller strike on a third. Turtles were otherwise healthy and active at the time of capture.

View Figure

• Download as Powerpoint
• Download Figure

At least 6 crab species, 9 mollusc species, tubes from 2 worm species, 1 echinoderm, and 1 bony fish were found in the fecal samples (Table 1). Over half (57%) of the samples contained common blue crab and/or Callinectes spp., and the next most commonly occurring animal food items > 1 cm were polycheate worm tubes (29%), hermit crabs (19%), calico box crab (Hepatus epheliticus; 10%), and mottled purse crab (10%). The sole occurrence of a moon snail shell (Neverita duplicata) was in a sample that also contained a hermit crab.

Table 1. Frequency of occurrence (%F) of items found in fecal samples from hook-and-line–caught and entangled Kemp's ridley sea turtles in Galveston County, Texas, United States, 2005–2008 (n = 21).

View Fullscreen

Macroaglae, Sargassum spp., occurred in 48% of samples (26.1–41.2 cm SCL), whereas 43% contained molluscs < 1 cm, mollusc fragments, shell hash, and/or rocks (“bottom material”, 26.1–44.5 cm SCL), and 19% contained terrestrial plants (33.7–41.2 cm SCL; Table 1). The most commonly occurring mollusc < 1 cm was the grass cerith (Bittium varium), which was found in 24% of samples and co-occurred with Sargassum spp. in 2 samples (30.2 and 41.2 cm SCL). Sargassum spp. and bottom material co-occurred in 29% of samples (26.1–41.2 cm SCL). No pelagic prey items co-occurred with Sargassum spp., although one larger individual (41.2 cm SCL) had consumed 2 potentially floating items—a bird feather and terrestrial plant matter—in addition to Sargassum spp., benthic prey, and part of an aluminum can. Based on known distributions of Sargassum spp. in the northwestern Gulf of Mexico (Moretzsohn et al. 2011; Lehman 2013), these Kemp's ridleys most likely ingested 2 commonly occurring floating species, S. fluitans and S. natans; however, a sessile species, S. pteropleuron, has also been documented offshore (60–200 m depth; Moretzsohn et al. 2011). Sargassum contained in samples were limited to small quantities of leaves and/or bladders, making species identification difficult. Of 6 samples containing Sargassum bladders, 4 were likely S. fluitans (no projections on bladders) and 2 were likely S. natans (projections on bladders). Four other samples contained only Sargassum leaf fragments, prohibiting species identification.

Anthropogenic debris was consumed by 2 Kemp's ridleys (candy wrapper, aluminum can), and mullet (Mugil sp.), a bony fish commonly used as recreational fishing bait, was consumed by another. No additional hooks or fishing gear were documented aside from those associated with the turtles' initial capture events, neither in fecal samples nor via radiography.

DISCUSSION

Although sample size was relatively small, crabs were clearly important components of the diet of Kemp's ridleys caught on recreational gear in Galveston County. The common blue crab, a portunid, was the most common crab species, whereas hermit crabs, calico box crabs, and mottled purse crab, all walking crabs, were also observed relatively frequently. Molluscs did not appear to be common target prey items, with only a sole Atlantic moon snail (shell likely inhabited by a hermit crab), 3 larger bivalve species, and various shell fragments and small snails ingested during the course of benthic foraging. Crab-rich diets have also been reported for Kemp's ridleys in New York (Burke et al. 1993, 1994), Virginia (Seney and Musick 2005), Georgia (Frick and Mason 1998), west-central Florida (Schmid 1998), the northwestern Gulf of Mexico (Werner 1994), and southern Texas (Shaver 1991).

Worm tubes, and presumably the polychaetes that inhabited them, were also targeted by some Kemp's ridleys, as has been observed elsewhere in the Gulf of Mexico. Large-scale studies in southwestern Florida (Witzell and Schmid 2005), Sabine Pass (Werner 1994), and southern Texas (Shaver 1991) found unidentified worm tubes in 41%, 21%, and 4% of samples, respectively. The south Texas study (Shaver 1991) also observed a 6% frequency of occurrence for the plumed worm (Diopatra cuprea), which was found in 29% of samples in the current study. The relatively high occurrence of worm tubes in the current study suggests that live-bottom habitat serves as an important foraging substrate, as seen in the Ten Thousand Islands of southwestern Florida (Schmid and Barichivich 2006).

Co-occurrence of Sargassum spp. and bottom material in one-third of the samples suggests that some level of surface feeding may continue for Kemp's ridleys beyond the 20–25-cm SCL size range. However, it is more likely that the Sargassum was consumed nearshore as benthic detritus, as evidenced by its co-occurrence with small benthic molluscs (e.g., grass cerith), shell hash, and rocks but not with pelagic prey items. Larger samples sizes are necessary to further elucidate dietary patterns, but recently recruited Kemp's ridleys shifting from pelagic foraging may initially consume worm tubes prior to focusing on crabs. Individuals may also be exhibiting opportunistic feeding, and turtles associated with piers and other hard structures that create “artificial live bottom” may be more likely to forage on worm tubes. The strategy of foraging on sessile prey prior to motile prey, however, is consistent with observations of small Kemp's ridleys in New York feeding on slower, walking crabs (Burke et al. 1993, 1994) and larger conspecifics further south feeding more often on faster, swimming crabs (Shaver 1991; Werner 1994; Seney and Musick 2005).

Coastal waters along the upper Texas coast are nutrient-rich and thereby able to support relatively large populations of the benthic crustaceans consumed by Kemp's ridleys (Metz 2004). Human activities may provide additional, opportunistic food sources beyond the local crab populations. Undoubtedly, hook-and-line caught Kemp's ridleys were attracted to baited hooks, but associations with fishing piers may have also been influenced by invertebrate communities living on or near the structural supports of the piers (e.g., Glasby and Connell 1999). Similar associations have been inferred for loggerhead sea turtles displaying fidelity to offshore oil platforms (Renaud and Carpenter 1994; Gitschlag et al. 1997), which support a host of invertebrates (Ellis et al. 1996). Unwanted bait and catch discarded by recreational anglers, as well as shrimping bycatch, may provide additional food sources (e.g., Shaver 1991). Fecal samples examined here, however, showed little evidence of this feeding behavior; only one sample contained fish, none contained squid or shrimp, and the one that contained a scavenging dog whelk, Nassarius vibex, included only other benthic prey. Feeding behaviors may change seasonally (e.g., Seney and Musick 2005), but turtles may also exhibit fidelity to piers once they begin to utilize them as a foraging area (e.g., Rudloe and Rudloe 2005). Kemp's ridleys sampled in this study were relocated upon release to minimize future interactions (Seney and Landry 2011), but one, which was released 32 km (20 mi) offshore, was recaptured at a nearby pier within the same season.

In addition to providing foraging opportunities, the upper Texas coast puts Kemp's ridleys at risk for direct interactions with human activities, as evidenced by fishing-gear interactions and samples containing anthropogenic debris and mullet. Such interactions have also been observed in southern Texas, where plastic, rubber, and fish were consumed by immature ridleys (Shaver 1991). None of the turtles sampled in the present study were caught on an artificial lure, and all but one were hooked, suggesting these individuals were targeting baited hooks as an opportunistic foraging behavior. Diet of benthic-stage Kemp's ridleys may also put the species at risk for indirect interactions with commercial fisheries. The preferred food item, common blue crab, is commercially exploited (Hammerschmidt et al. 1998), and it may also be taken as bycatch in other fisheries (e.g., Scott-Denton et al. 2012). Declines in blue crab populations and other natural prey items could precipitate shifts to anthropogenic food sources, such as baited hooks, discarded shrimp bycatch, and actively fishing nets and traps (Shaver 1991; Seney and Musick 2007). Such diet shifts would expose Kemp's ridleys to increased direct interactions with fisheries and, thereby, to hooking, entanglement, and drowning risks.

The impacts of anthropogenic factors, such as debris ingestion and interactions with recreational fishing, within the upper Texas coast developmental habitat should be reviewed by managers to ensure the Kemp's ridley's recovery. Such an assessment is particularly important in light of the recent, abrupt change in the species' nesting trends in 2010 (NMFS and US FWS 2015). Further characterization of diet will allow for better understanding of Kemp's ridley ecology, any temporal diet variation, and potential anthropogenic interactions and competition, ultimately aiding in species management and recovery.