Introduction

Hematophagous bats are endemic to Latin America and are represented by three species: Desmodus rotundus, Diaemus youngi, and Diphylla ecaudata. The common vampire bat (D. rotundus, hereinafter referred to as vampire bat) is the most widely distributed blood-sucking species in America (^{Barquez et al. 2018}) and the only one that preys on wild mammals (^{Sánchez-Cordero et al. 2011}; ^{López-Tello and Mandujano Rodríguez 2016}; ^{Pignaton Gnocchi and Srbek-Araujo 2017}). However, when the availability of domestic species increases, the vampire bat tends to feed on them more frequently, with an apparent preference for horses, boars, and poultry (^{Mialhe 2014}; ^{Bobrowiec et al. 2015}).

Changes in the availability and abundance of wild species, coupled with the intensification and expansion of livestock rising, have contributed to the modification of the feeding habits of the vampire bat, resulting in the growth and geographic expansion of its populations (^{Streicker and Allgeier 2016}). The area around Laguna de Términos, Campeche, is a highly fragmented region whose natural areas have been transformed into agricultural and livestock-raising land (^{Soto-Galera et al. 2010}). This region is home to one of the highest livestock densities in the state (^{INEGI 2007}; ^{Robinson et al. 2014}), an ideal feature for the establishment of vampire bat populations. The establishment of these populations may trigger economic and public-health issues, because vampire bats are vectors of zoonotic infectious diseases with the potential to infect wild and domestic animals, and even humans (^{Mayen 2003}; ^{da Rosa et al. 2006}; ^{Young et al. 2017}). Such effects can be exacerbated by the availability of abundant food resources, as those derived from the introduction of invasive exotic species such as the European Wild pig (Sus scrofa) and its hybrid variety, the wild pig (^{Galetti et al. 2016}).

The presence of wild pigs in the area adjacent to Laguna de Términos was reported since 2014 (^{Hidalgo-Mihart et al. 2014}). However, it is estimated that the species was introduced for hunting purposes since the early 1990s. Its distribution and abundance have increase since then (^{Hernández-Pérez et al. 2018}), in parallel with the density of cattle and horses as reported by local farmers. The increasing abundance of both livestock and wild pigs in the region means higher prey availability for the vampire bat, probably fostering the growth of their local populations (^{Delpietro et al. 1992}).

The vampire bat is considered as an obligate hematophagous species, as it feeds exclusively on blood (^{Mayen 2003}). There are few studies that accurately describe the type of prey attacked by this bat and its tendency to feed on domestic species (^{Mialhe 2014}; ^{Bobrowiec et al. 2015}; ^{Streicker and Allgeier 2016}). A useful tool to determine the type of prey attacked by the vampire bat, in addition to molecular studies, has been the recording of observations both directly and through camera traps (^{Galetti et al. 2016}; ^{López-Tello and Mandujano Rodríguez 2016}; ^{Pereira et al. 2016}; ^{Pignaton Gnocchi and Srbek-Araujo 2017}). This work reports the first evidence for Mexico regarding vampire bats attacking wild pigs in Laguna de Términos, Campeche, one of the most important wetlands in Mexico.

Methods

Photographs were captured with camera traps in the area of influence of the Laguna de Términos Flora and Fauna Protection Area (APFFLT), Campeche, in southern Mexico (Figure 1). Camera traps were installed at 50 cm from the ground on trees adjacent to trails where we found traces of wild pigs and collared peccaries, as part of the project “Distribución y relaciones ecológicas entre cerdos ferales (Sus scrofa) y poblaciones nativas de pecaríes” (^{Hernández-Pérez 2019}). A total of 11 camera traps were placed (Cuddeback® Black Flash E3) in five localities, with a distance of at least 500 m between traps. Camera traps were set to operate 24 hours a day and to capture a series of five photographs and one 20-second video per detection event. The overall photo trapping period considering the five sites took place from 25 January to 16 August 2018, involving a sampling effort of 1,428 trap days. Vampire bats were identified based on the characteristics that allow the species to adopt a quadruped position to land and move on its prey, including the long thumb with pads on its surface, movement lifting the body, and small jumps on the back of its prey (^{Pereira et al. 2016}), and even the modified nasal sheet, typical of the species. A photograph or video of an individual in a camera trap was considered as an independent record when it was separated from the next record in the same camera trap by at least 60 minutes; if more than one individual is viewed in the same photograph or video, then each of these individuals was considered as an independent record. The activity pattern of vampire bats and other species that had at least ten independent records caught in the same camera as the interaction between wild pigs and bats, was evaluated by circular descriptive statistics with the software Oriana 4.0 (^{Kovach Computing Service 2011}).

Figure 1 Location of the Laguna de Términos region in southern Mexico. Black dots mark the location of camera traps; the black triangle marks the site where the interaction between wild pigs and vampire bats was recorded. 

Results

A total of 18 species of mammals were recorded in the 11 camera traps installed in the five locations (Table 1). Wild pigs were recorded separately on 213 occasions in six camera traps placed at three different locations.

Predation of the vampire bat on wild pigs was recorded in a single camera trap (18.360464° N, -91.269592° W) installed in an oilpalm (Elaeis guineensis) plantation surrounded by livestock pastures and patches of secondary tropical forest around the village of Ignacio Zaragoza, municipality of Carmen, Campeche (Figure 1). The photo trampling periods for this camera trap ranged from 27 February to 16 April, and from 27 April to 16 August 2018, with a sampling effort of 162 trap days.

This same camera trap captured 486 independent records of seven species: Sus scrofa (n = 188), Bos sp. (n = 134), Procyon lotor (n = 87), Canis latrans (n = 42), Equus ferus (n = 23), Didelphis virginiana (n = 11), and Sylvilagus floridanus (n = 1). Wild pigs were the only species recorded as prey of the vampire bat. Of all the independent records of wild pigs, 11.7 % (n = 22) are a series of photographs and videos of the interaction between wild pigs and vampire bats (Figure 2). Most of the records show foraging wild pigs, primarily adult males (n = 7), carrying a vampire bat either on one side of the back or on the neck, just behind the ears. Two video records captured a bat flying and landing directly on the lumbar portion (posterior dorsum) of the prey, then moving toward the front. A separate event recorded two vampire bats flying over two juvenil wild pigs. However, there were no records of predation to juvenil pigs.

Figure 2 Photographic record of vampire bats (Desmodus rotundus) attacking wild pigs (Sus scrofa) in the Laguna de Términos area, Campeche, Mexico. 

The species Bos sp., Canis latrans, and Equus ferus showed a cathemeral activity pattern, while Sus scrofa, Procyon lotor, and Didelphis virginiana were more active at night (Figure 3). Attacks of vampire bats on wild pigs took place at night, between 20:00 and 5:00 h, with a peak of activity around 23:21 h (Figure 3).

Figure 3 Pattern of activity of a) Bos sp. b) Canis latrans, c) Equus ferus, d) Sus scrofa, e) Procyon lotor, and f) Didelphis virginiana, compared to g) the period when vampire bats (Desmodus rotundus) attacked wild pigs (Sus scrofa) in the Laguna de Términos area, Campeche, Mexico. Bars show the percentage of independent records. The vector represents the 95 % confidence interval. 

Discussion

The predator-prey interaction between vampire bats and wild pigs had only been reported in Brazil (^{Galetti et al. 2016}; ^{Pereira et al. 2016}). This work is the first to report evidence of this interaction in Mexico. In spite of the presence of wild species in the study area, including the collared peccary (Pecari tajacu) and the white-tailed deer (Odocoileus virginianus), as well as domestic species such as bovines (Bos sp.) and horses (E. ferus), no attacks on these species were recorded, all of which are also preyed by the vampire bat (^{Mialhe 2014}; ^{Bobrowiec et al. 2015}). In interviews to local farmers, no attacks of bats on cows or horses were documented during the period monitored.

^{Bobrowiec et al. (2015)} suggest that the vampire bat tends to feed on domestic pigs when these are easily available. This trend may be related to the activity pattern of pigs, the thickness of the skin, and the presence of a marginal vein in the ears from which the vampire bat can suck sufficient blood for several days. This work documents that wild pigs are attacked by vampire bats most frequently in the Laguna de Términos area. Our findings suggest that vampire bats are currently feeding preferentially on wild pigs due to the expansion of this exotic prey in the study area in recent years. An alternative explanation is that, unlike the other species recorded, wild pigs are active mainly at night, matching the activity pattern of vampire bats (Figure 3). This case implies that the vampire bat would be behaving opportunistically as a non-selective predator, feeding on the most abundant prey species during its period of activity.

The potential distribution for wild pigs as estimated in the study area suggests the expansion of their populations to new sites as a result of the great ecological plasticity of this species, but also of other socioeconomic factors that are promoting the conversion of natural vegetation to oilpalm crops and induced pastures (^{Hernández-Pérez 2019}). The increased availability of wild pigs reported by farmers, together with changes of land use and extensive livestock ranching, constitute a potential scenario for the establishment, growth and expansion of local populations of vampire bats (^{Pedrosa et al. 2015}). This increase may transform the region into a high-risk area for the transmission of zoonotic diseases such as rabies (^{Gavier-Widén et al. 2012}; ^{Young et al. 2017}). The rabies virus has been identified in vampire bat populations thriving in the municipalities around Laguna de Términos. The municipality of Candelaria is worth mentioning, as it is currently considered an endemic zone of the disease (^{SENASICA 2016}; ^{SENASICA 2017}). However, epidemiological data are limited to reports of positive cases. Under the current context described above, in-depth ecological and epidemiological studies addressing the interaction between wild pigs and vampire bats are needed to assess the risk of transmission of diseases to other wild species and to humans in the region. Wild pigs move freely across the region between areas covered by natural vegetation and anthropized landscapes (^{Hernández-Pérez 2019}), with numbers that will certainly increase in the future based on their high reproductive rate, as a single female can give birth to between eight and 14 pups per year (^{Taylor et al. 1998}). Immediate control measures are required at least in the Laguna de Términos Flora and Fauna Protection Area to manage the expansion and establishment of wild pigs as well as their negative effects on the regional livestock activities, human health, and the conservation of wild populations.