Feeding habits of the jaguar Panthera onca (Carnivora: Felidae) in Tortuguero National Park, Costa Rica

ISSN-0034-7744

The most commonly used technique of these studies has been analysis of scat samples, with other methods such as examination of carcasses used less frequently.These noninvasive techniques can be of great value when studying cryptic and endangered species such as the jaguar, since it is not necessary to capture the animal (Waits & Paetkau, 2005;Perilli, Lima, Rodrigues, & Cavalcanti, 2016).Klare, Kambler and Macdonald (2011) also emphasized that non-invasive techniques are relatively simple and affordable compared with other methods (e.g.stomach content).However, some disadvantages such as the low sample size and poor detectability of prey remains, must be considered to avoid biased or spurious results (Foster, Harmsen, & Doncaster, 2010;Tambling et al., 2012;Steenweg, Gillingham, Parker, & Heard, 2015;Morin et al., 2016).
The present study expands the current knowledge of jaguar diet in Tortuguero National Park, Costa Rica.Previous research in this area has focused on the predator-prey interaction between the jaguar and the local nesting population of sea turtles (Veríssimo et al., 2012;Arroyo-Arce & Salom-Pérez, 2015).Although these studies highlighted the importance of the green sea turtle (Chelonia mydas) as a food source for the local jaguar population, the role of other prey species is still unknown.

Study site:
The study area was defined by approximately 29 km of coastline located within Tortuguero National Park, Costa Rica (10 o 32'28'' N -83 o 30'08'' W).This site, known as Tortuguero beach, extends from the Tortuguero River mouth in the North to the Jalova River mouth in the South; there is also a human-made trail running parallel and along the entirety of the beach (Fig. 1).A Tropical Wet Forest (Holdridge, 1969) borders the entire study site except for a coconut plantation and cattle ranch to the South, and Tortuguero village to the North end.The average temperature ranges from 25 to 30 o C, with a mean annual precipitation of 6 000 mm (Bermúdez & Hernández, 2004).

Data collection and analysis:
Carcass data of potential prey species, believed to have been predated by a jaguar, was collected by carrying out weekly surveys along Tortuguero beach, from July 2012 to November 2016.Additional information was also collected opportunistically while the researchers surveyed the human-made trails surrounding the study area.When a carcass was discovered, the carcass was examined for evidence of jaguar predation (e.g.bite marks on the neck, drag marks).If determined that a jaguar was responsible for the kill, the following information was recorded: date of discovery, geographic coordinates and species identification of the carcass (for more details on the methodology see Arroyo-Arce & Salom-Pérez, 2015).
Jaguar scats were collected, while performing the carcass data collection, along the beach and trails, from August 2012 to December 2016.Upon discovery, the sample was collected in a plastic bag labeled with the date of collection and geographic coordinates.Jaguar scats were differentiated from other wild cats and carnivores based on their morphological characteristics (Aranda, 2000) and associated signs (e.g.nearby jaguar tracks).It is important to highlight that samples were not collected when it was not possible to say with certainty that they belonged to jaguars.
Each scat was soaked in a solution of water and detergent for up to 48 hours.After this period, the scat was washed with running water over fine-mesh filters and sun dried.Once dried, the non-digested prey remains (e.g.claws, tissue, hair) were separated and later identified to the lowest possible taxonomic level using specialized guides (Baca & Sánchez-Cordero, 2004;Pech-Canché, Sosa-Escalante, & Koyoc, 2009;Juárez et al., 2010) and specimens from scientific collections.Bones, arthropod fragments and plant material found in the samples were not identified.
The frequency of occurrence (FO) and percent of occurrence (PO) were used to describe the scat contents.FO refers to the percentage of fecal samples in which a prey species was found (Equation A), while PO refers to the number of times a specific item was found as percentage of all items found (Equation B; Ackerman, Lindzey, & Hemker, 1984).
Where fi= number of fecal samples in which species i was recorded, N= total number of fecal samples.

Camera trap records (photos or videos)
clearly showing jaguar preying on terrestrial preys, or prey species with evident marks of a jaguar attack (e.g.claw or bite marks) were used as complementary information.These records were obtained as part of the long-term camera trap monitoring program conducted by Coastal Jaguar Conservation in Tortuguero National Park, and corresponded to 12 camera trap stations located along the human-made trail running parallel to the beach from June 2012 to December 2016.

RESULTS
Of 1 374 carcasses recorded (mean per year: 276.8 ± 84.12 SD), 99 % corresponded to sea turtles, the green sea turtle being the most common (Table 1).
Of 75 scats analyzed (mean per year: 12.50 ± 8.38 SD), 47 % consisted of a single prey item, 17 % had two or more prey items, and prey species were unidentifiable in 36 % of the samples.The white-lipped peccary was the most frequent prey found on the scat analysis.Furthermore, one scat collected on February 06, 2014 (10°22'374'' N -83°81'158'' W) contained seven claws that, due to their morphology, could belong to an adult jaguar or a puma (Puma concolor) (Table 2).Another important record was the presence of the remains of a green sea turtle hatchling in one scat collected on December 05, 2016 (10°28'614'' N -83°28'158'' W).
As for the camera trap records, a total of two photos and one video, evidenced jaguar predation upon terrestrial prey species, including the nine-banded armadillo (Dasypus novemcintus), the Baird's Tapir (Tapirus bairdii) and the common opossum (Didelphis marsupialis), respectively (Table 1).

DISCUSSION
Our results were comparable with those obtained in other studies in Costa Rica (Chinchilla, 1997;Carrillo et al., 2009;Montalvo, 2012), where sea turtles and ungulate species were considered major prey species for the jaguar.In Tortuguero beach, the prey selectivity towards the green sea turtle was expected as this species is highly abundant and available throughout most of the year (Arroyo-Arce & Salom-Pérez, 2015).Although abundance of white-lipped peccary has not yet been estimated, important records of groups of more than 20 individuals have been reported roaming within the study area (Coastal Jaguar Conservation, unpublished) year-round.Our study also highlights the variety of species (e.g.mammals, birds, reptiles) forming the jaguar's diet, reflecting its well-known generalist and opportunistic behavior (Rabinowitz & Nottingham, 1986).
The remains of a large feline in one of the scats are of great interest, since predation among big cats has been rarely documented.For example, in Brazil jaguar infanticide (Soares et al., 2006;Tortato et al., 2017) and jaguar cannibalism (Azevedo, Costa, Concone, Pires-Da Silva, & Verdade, 2010) has been reported.Regarding pumas, Cassaigne et al. (2016) documented an un-collared male killing and eating a collared female puma in Sonora, Mexico.Since Tortuguero beach has a high density of jaguars (yearly average of 13 individuals detected along 29 km of coastline; Coastal Jaguar Conservation, unpublished data), perhaps a jaguar attack upon a conspecific or a puma could have been triggered by social stress caused by multiple individuals converging in one spot or, a female defending her off-spring, or individuals defending their prey (Azevedo et al., 2010).
A major limitation of carcass analysis was the low detectability of prey remains in closed habitats, which could lead to an underrepresentation of some prey species; especially smaller ones (Tambling et al., 2012).In our study, all carcasses were recorded along the beach, which being considered an open habitat facilitated the easy detection of carcasses, and belonged to both small (≤ 1 kg; n= 2) and large (≥ 15 kg; n= 1 370) prey species.Other factors that facilitated the detection were presence of black vultures (Coragyps atratus) around the kills and the large size of sea turtle remains (sea turtle carapace length ranges from 81-183 cm, which varies by species).Carcasses were not detected along the trail (closed habitat), as a result of the remains being obscured by the vegetation, thus hindering detection by researchers.
Apart from green sea turtle hatchling, no other evidence of sea turtles was discovered in scat samples, despite of being one of the main components in the jaguar's diet in Tortuguero National Park.The under representation of this item could imply that jaguars only feed on sea turtles' soft tissues (e.g.internal organs), leaving intact the hard parts (e.g.carapace, head, flippers).Since consumed parts are easily decomposed within the digestive system, they are unexpected in scat samples (Hayward et al., 2016).
Field sample misidentification is another disadvantage of scat analysis.Some authors (Farrell, Roman, & Sunquist, 2000;Davison, Birks, Brookes, Braithwaite, & Messenger, 2002;Morin et al., 2016) suggest that identification based on sample morphology (e.g.size, shape, dietary content) may not be the most reliable way to determine the species of origin, specifically between carnivore species of similar size.In Tortuguero National Park, jaguar scats were identified based on the appearance but also on other signs (e.g.presence of jaguar tracks near by the samples).Furthermore, jaguar presence is frequent, while pumas are scarce in the study site (Arroyo-Arce & Salom-Pérez, 2014).Therefore, we are prone to believe that our sampling identification in the field was accurate.However, we recommend that future studies use molecular analysis of scats to minimize misidentification (Dalén, Gotherström, & Angerbjörn, 2004).Both carcass and scat analyses underestimated the importance of domestic animals as part of the jaguar diet.Although we did not find carcasses of domestic animals or their remains in scat content, we are aware of several confirmed jaguar attacks upon domestic cows (Bos taurus) and dogs (Canis familiaris) in the cattle ranch (Southern end of the beach), the ranger station and communities located around the study area.Therefore, we recommend the use of other methods (e.g.interviews, molecular analysis) to determine the role that domestic animals play on the feeding habits of the local jaguar population.This is the first broad study of jaguar diet in Tortuguero National Park.Our results not only provide additional information regarding well-known predator-prey relationship between sea turtles and the jaguar, but also to understand the role that other prey species play in jaguar diet.Our results also highlight the importance of combining different methods to have a better representation of the jaguar feeding habits.Further research to complement our study should focus on calculating terrestrial prey species abundance in the study area, and collecting samples from other locations (e.g.further inland), information that would help to complement our study.

TABLE 1 Jaguar
Panthera onca dietary composition from carcasses and camera trap records documented in the coastal habitat of Tortuguero National Park, Costa Rica.Period: June 2012-December 2016 Camera trap record of a jaguar holding a prey in its mouth.2.Camera trap record of an individual with jaguar bite marks alongside its body.

TABLE 2 Jaguar
Panthera onca dietary composition from scats (n= 75) collected in the coastal habitat of Tortuguero National Park, Costa Rica.Period: August 2012-December 2016 One scat contained seven claws of Panthera onca or Puma concolor.2. Remains of Bradypus variegatus and/or Choloepus hoffmanni.3.One scat with remains of a Chelonia mydas hatchling.