Cannibalistic necrophagy in red foxes: do the nutritional benefits offset the potential costs of disease transmission?

Cannibalism, or intraspecific predation, occurs when an individual consumes another individual of its own species; c.f. ‘cannibalistic/conspecific/intraspecific necrophagy’ when the individual consumes all or part of a dead conspecific. These behaviors are widespread across animal taxa but are considered rare among mammalian carnivores. The consumption of conspecifics could involve ecological and nutritional benefits in providing high-quality resources. However, it can carry potential risks such as those related to pathogen transmission. Despite this, the overall role of conspecific consumption in disease transmission among animals has received little attention. Here, we report the first direct observations of cannibalistic necrophagy in red foxes Vulpes vulpes. Our photographs of this behavior prompt speculation on the role of cannibalism in fox population dynamics and intraspecific disease transmission. We placed six fox and seven lamb carcasses of similar size to determine possible differences in the scavenging behavior of red foxes between dead conspecifics and heterospecifics. Conspecific necrophagy was recorded at all fox carcasses, involving foxes scavenging fox carcasses at different stages of decay. Carcass detection time by foxes was similar at both fox and lamb carcasses, but lambs were completely consumed in a shorter period, evidencing a preference for heterospecific consumption. Our findings contrast with other studies which argue that cannibalism avoidance in mammalian carnivores is due to an evolutionary strategy to reduce the probability of disease transmission. In fact, our observations of conspecific necrophagy provide direct evidence against the parasite-avoidance hypothesis, suggesting that carnivore and conspecific carcasses can represent an alternative trophic resource for foxes in certain areas and circumstances.

behaviors have been observed in a wide variety of animal taxa, involving species of different trophic positions, including carnivores (Fox 1975;Polis 1981;Elgar and Crespi 1992;Fouilloux et al. 2019). Feeding on conspecifics is generally considered an adaptive strategy involving both ecological and nutritional benefits, reducing competition and providing high-quality resources (Snyder et al. 2000;Mayntz and Toft 2006). However, this practice also carries several potential costs such as the risk of injury and death during a predatory attack, the loss of inclusive fitness when a relative is eaten, and in particular, the risks associated with pathogen transmission (i.e. ingestion of bacteria, viruses, prions, parasites or fungi ;Boots 1998;Pfennig and Hoffman 1998;Rudolf and Antonovics 2007) which increase when feeding on phylogenetically similar species (Pfennig 2000).
Cannibalism can occur under various stressful conditions, including food scarcity (quantity/quality), high conspecific density, sexual conflict (e.g. when females consume males if food is limited, and when unrelated offspring are eaten) and in response to a variety of abiotic conditions (Polis et al. 1984;Fouilloux et al. 2019). It is a widespread phenomenon in nature and has been documented in at least 18 carnivore mammal species (Bertram 1975;Polis 1981;Amstrup et al. 2006;Azevedo et al. 2010;van Neer et al. 2019). However, direct observations of cannibalism by carnivore mammals are unusual, being in most cases based on evidence of infanticide events or suggested by non-conclusive indirect signs (diet analyses) (Polis et al. 1984;Ebensperger 1998;Remonti et al. 2005). Direct observations of carnivores scavenging on carnivore carcasses are less common, especially involving conspecifics (Abernethy et al. 2016;Moleón et al. 2017;Muñoz-Lozano et al. 2019). Here, we report the first direct observations of conspecific necrophagy in red foxes Vulpes vulpes (Table 1, Fig. 1) and our photographs of this behavior prompt speculation on the role of necrophagy in fox population dynamics and intraspecific disease transmission.
Six red fox (mean weight ± SD: 5.9 ± 0.96 kg) and seven lamb carcasses (6.7 ± 1.23 kg) were monitored in the Pyrenees (N Spain) using motion-triggered remote cameras from December 2017 to August 2019. Carcasses were collected as road kills (foxes) or obtained from legal farms (lambs) and were placed as entire corpses in accordance with all national and international legal requirements. Fox carcasses were frozen (− 20 °C) in individual bags prior to carcass placement. Cameras were located close to the carcasses (5-10 m away) and were programed to take a set of three consecutive photos when triggered by motion, with one second delay between each photo, and wait for 15 s before taking the next set of photos. Cameras were operational 24 h a day and were visited every 1-4 days to check them and to examine the general state of the carcass (e.g. degree of decomposition), which was also checked from the photos. Carcasses were placed during daylight hours at altitudes ranging from 700-1500 m asl in various seasons and landscape types: 'open', without shrub vegetation ( Fig. 1A-D); 'semi-closed', with abundant shrub cover (Fig. 1E); and 'closed', within forests of Scots pine Pinus sylvestris (Fig. 1F) (Table 1). Conspecific necrophagy was recorded at all fox carcasses, and foxes completely scavenged two fresh carcasses, one actively decaying and two in an advanced state of decay, while in one carcass in advanced decay only bone remains were consumed (see Matuszewski et al. 2008 for details on carrion decomposition stages) (Table 1, Fig. 1). Fox carcasses were also consumed by other facultative avian and mammalian scavengers and in one case by an obligate scavenger, the griffon vulture Gyps fulvus (Table 1). In contrast, foxes were almost entirely responsible for consuming the lamb carcasses (although in two cases consumption by wild boar was also recorded). Table 1 Main variables related to carcass characteristics and scavenging dynamics (scavenger species richness and total consumption time) and those related to carcass consumption by foxes (carcass detection time and start of consumption in relation to carcass placement) 'Area' refers to landscape characteristics: "O" (open), "S" (semi-closed) and "C" (closed). Numbers in parentheses show the number of feeding events by each species. 'State of carcass decay' refers to the carcass decomposition stage at the start of consumption by foxes Carcass detection time by foxes was similar for both fox and lamb carcasses (mean ± SD: 81.44 ± 72.34 vs. 50.16 ± 34.68 h respectively, Mann-Whitney U test, p = 0.72). In contrast, the 'consumption time' (time elapsed between carcass placement and its complete consumption) of fox carcasses was longer than lamb carcasses (420.03 ± 181.27 vs. 86.55 ± 19.25 h respectively, Mann-Whitney U test, p = 0.0034) (Table 1). Accordingly, a higher number of fox feeding events, defined as consecutive photos of individuals of different species or consecutive photos of individuals of the same species made more than 30 min apart (O'Brien et al. 2003), were recorded at fox carcasses compared to lamb carcasses (20.50 vs. 5.29 fox feeding events per carcass) (Table 1). Fox carcasses showed a significantly higher species 'richness' (number of scavenger species per carcass) compared to lambs (3.5 ± 1.37 vs. 1.29 ± 0.49 species per carcass respectively, Mann-Whitney U test, p = 0.006) since all lamb carcasses were completely consumed in a shorter period after detection (Table 1).
Red fox cannibalism has only been suggested through indirect observations in Italy (Pozio and Gradoni 1981;Rossi et al. 1992), but the frequency of occurrence of potential cannibalism events in the NW Italian Alps was rare (Remonti et al. 2005). Because these results were based on indirect evidence (fecal analyses and stomach contents), it is possible that the inferred cannibalism was confused with coat-cleaning (i.e. ingesting some of their own hair which pass almost unaltered through the digestive tract), infanticide or the ingestion of dead cubs (Frank 1979). In addition, Cagnacci et al. (2003) analyzed 922 Italian red fox scats and found no evidence of cannibalism.
Our direct observations of red fox conspecific necrophagy contrast with other studies such as that of Muñoz-Lozano et al. (2019) who monitored 20 red fox carcasses in two mountainous Mediterranean areas in SE Spain. These authors only confirmed consumption of fox carrion by golden eagles Aquila chrysaetos, although the carcasses were frequently visited by mammalian carnivores including red foxes. Another study in the same area monitored 46 carnivore carcasses (including 36 foxes) and found that most mammals recorded visiting carnivore carcasses avoided consuming them. Only two carcasses were consumed by foxes and one by a wild boar, but the carcasses were always of heterospecifics, and conspecific consumption was avoided  (Moleón et al. 2017). In addition, Díaz-Ruiz et al. (2013) reviewed 55 studies on fox diet in the Iberian Peninsula and found no evidence of cannibalism. These results suggest that the cannibalistic necrophagy behavior we observed is indeed unusual and might relate to the particular characteristics of our study area, perhaps related to trophic availability. Necrophagy could bring benefits compared to predation; fresh carrion provides nutritional value equal to predated food, but without the risks and energy expenditure associated with hunting (DeVault et al. 2003). However, regarding cannibalism as a possible disease transmission route, the review by Rudolf and Antonovics (2007) showed that intraspecific predation was the predominant mode of disease transmission in only a few cases and species, while necrophagy or heterospecific trophic transmission were more important mechanisms of disease transmission in several species. In fact, necrophagy was shown to be responsible for the transmission of multiple nematodes of the genus Trichinella affecting carnivores (Pozio 2000). This is especially relevant in areas in which species with highly opportunistic foraging habits such as foxes could show higher scavenging tendencies when carrion is relatively abundant, for example in mountainous areas with high carcass availability (Cagnacci et al. 2003;Remonti et al. 2005).
Regarding scavenging dynamics, we found that carcass detection time by foxes was similar for both fox and lamb carcasses, although lambs were completely consumed much more quickly. In contrast, the consumption time of fox carcasses was nearly five times longer than for lambs and they were visited by more scavenger species, including facultative and obligate scavengers (Table 1). Despite the idea that carnivore carrion is avoided by mammalian scavengers, we found that carnivore carcasses such as foxes could become alternative trophic resources in certain areas. However, the more rapid consumption of lamb carcasses by foxes compared to fox carcasses confirms that foxes prefer carrion from heterospecifics over conspecific necrophagy. Furthermore, foxes took longer to commence consumption of fox carcasses than lambs after detection (224.64 ± 162.35 vs. 7.4 ± 12.62 h respectively, Mann-Whitney U test, p = 0.023) (Table 1), supporting a preference for heterospecific consumption.
The consumption of conspecifics by foxes may influence the transmission of meat-borne diseases such as trichinellosis, one of the most widespread helminthic zoonosis in nature. Current literature suggests that one of the main factors contributing to the maintenance of trichinellosis is related to carnivore species with cannibalistic and scavenging behaviors (Pozio 2000). In fact, the consumption of carrion is considered the main mechanism of transmission (Campbell 1988;Rossi et al. 1992). Of the three recognized Trichinella species, T. spiralis and T. britovi are the most widespread in wild and domestic animals in Europe (Pozzio 1998). In Spain, these two species are perpetuated in both domestic and sylvatic cycles (Rodriguez et al. 2008); T. spiralis is found in pigs, wild boars and foxes (Serrano et al. 1998;Criado-Fornelio et al. 2000) whereas T. britovi mainly infects wild animals, foxes being their main sylvatic hosts (Perez-Martin et al. 2000). Although the observed prevalence of T. britovi in foxes in the Pyrenees was low (0.3%), it was concentrated in mountainous and rural areas, reaching high local prevalences (López-Olvera et al. 2011). Globally, the prevalence of T. britovi has been correlated with the low availability of anthropic food resources in mountainous and sparsely populated areas (Pozio 1998). For instance, Rossi et al. (1992) suggested that cannibalistic behavior was more common in mountain than lowland foxes. Although these evidences were again based on indirect observations, red fox cannibalism was attributed to low food resources in mountainous areas, resulting in high Trichinella spp. prevalence in these areas. This agrees with our observations, since the six monitored carcasses scavenged by foxes were placed in low productive temperate mountainous areas with low human population density.
The most common hypothesis regarding cannibalism avoidance in carnivores attributes this behavior to an evolutionary strategy to reduce the probability of disease transmission (Pfennig and Hoffman 1998;Moleón et al. 2017) as evidenced by the scarcity of observations of conspecific necrophagy by mammalian carnivores. However, the ecological and evolutionary effects of cannibalism on parasite transmission deserve further consideration. Existing evidence suggests that cannibalism may be more likely to reduce parasite prevalence (e.g. by killing parasites in infected victims and reducing the number of susceptible hosts) (Van Allen et al. 2017). Nonetheless, there is clear empirical evidence that some wildlife diseases can be spread through cannibalism (Boots 1998;Pfennig 2000), especially when more than one individual consumes an infected victim (i.e. "group cannibalism", Rudolf and Antonovics 2007). Consequently, the idea that cannibalism spreads disease or significantly increases infection risk continues to dominate the ecological literature (Van Allen et al. 2017). Despite this, there have been no general studies on the overall role of cannibalism in disease transmission (but see Rudolf and Antonovics 2007;. Filling this gap would be worthwhile and future research should explore the role that cannibalistic behaviors might play in disease transmission in animals, to better understand the interactions between disease and conspecific consumption and to focus on the possible factors that drive this paradoxical behavior, especially those related to food availability, competitor removal, conspecific density, genetic predisposition, and local specializations. Our findings of conspecific necrophagy recorded at six fox carcasses by different individuals, do not support the parasite-avoidance hypothesis. Instead, they suggest that other factors, such as food shortages, could modulate conspecific necrophagy in mammalian carnivores, particularly foxes. Our observations of fox conspecific necrophagy in different years, seasons and areas (at carcasses ranging from 1.2 to 50 km apart) suggest that it is not an isolated occurrence. Furthermore, two foxes were simultaneously recorded completely consuming a single fresh fox carcass from the moment it was detected ( Fig. 1A-C). However, in contrast to lamb carcasses, most foxes only started to consume fox carcasses some days after detecting them, suggesting a preference for heterospecific carrion. This could also relate to differing individual personalities, since it is difficult to determine whether the individual fox detecting a carcass and not consuming it was the same one that ate it during the following days, implying that some individuals may avoid conspecific consumption, while others may not show this conspecific carrion avoidance trait. Because both the risk of parasite transmission and the nutritional quality of prey may decline with decreasing phylogenetic similarity between consumer and prey (Pfennig 2000), the interaction between these factors remains uncertain, suggesting that, in some circumstances, the increased nutritional benefit of cannibalism offsets the increased risk of pathogen transmission.
The role of cannibalism in red foxes and other mammalian carnivores is an important aspect of their ecology, but also has important implications for human health. Foxes could maintain and transmit zoonotic diseases to domestic animals, particularly those requiring multi-sectoral One Health approaches such as trichinellosis, which could maintain its prevalence through cannibalistic behaviors.
Author contributions POV, JT and AM conceived and designed the study. POV conducted the fieldwork. POV analyzed the data. POV, JT and AM wrote the manuscript.

Funding
This study has been possible as a result of the RTI2018-099609-B-C22 project, from the I + D + I National Plan funded by the Ministry of Science, Innovation and Universities.
Data availability Empirical data used for this research appear in the manuscript. Other additional data (photographs) can be requested to the authors.

Conflict of interest Authors declare no conflicts of interests.
Ethical approval All the work was conducted in accordance with relevant national and international guidelines, and conforms to all legal requirements in compliance with the Ethical Principles in Animal Research. Thus, protocols were conducted in accordance to the Regulation EC 1069/2009 of the European Parliament and of the Council of 21 October 2009 laying down health rules as regards animal byproducts and derived products not intended for human consumption, and in accordance to the guidelines approved by the Catalunya Government following the Law 42/2007, of 13 December 2008, on Natural Heritage and Biodiversity, the Legislative Decree 2/2008, of 15 April, approving the revised text of the Animal Protection Act and the Decree 148/1992, of 9 June, regulating the photographic, scientific and sports activities that may affect wildlife species. All experimental protocols were approved by the Catalunya Government (authorization reference SF/002).

Consent to participate Not applicable.
Consent for publication Not applicable.