The IUCN/SSC Canid Specialist Group's
African Wild Dog Status Survey and Action Plan (1997)

Chapter 4
Past and Future Causes of Wild Dogs' Population Decline

by Rosie Woodroffe & Joshua R. Ginsberg

In the previous chapter we showed how wild dog populations have been extirpated across much of Africa over the last 30 years. This chapter reviews the factors that might cause the few remaining populations to decline or disappear altogether:

Habitat fragmentation, persecution and loss of prey were the major causes of wild dogs' historic decline, and these factors still represent the principal threats today.

Competition with larger carnivores keeps wild dogs' numbers low, so that even the largest habitat fragments may contain populations too small to be viable.

Contact with human activity is directly responsible for over 60% of recorded adult mortality through road casualties, persecution and snaring. Even wild dogs living in large protected areas may stray over reserve borders where they are threatened by human activities.

Disease represents another serious threat to wild dogs, which has already caused the extinction of one population. The presence of people dramatically increases the disease risk to wild dogs, because domestic dogs provide a reservoir host for canid diseases.

As a result of these pressures:

• All of the wild dog populations remaining in Africa are under threat.
• In the long term, wild dogs living outside protected areas are unlikely to co-exist with growing human populations without innovative management.
• Even in large protected areas, wild dogs' long-term survival will depend on reducing potentially fatal contact with people and domestic dogs on reserve borders.

• Background
• 'Natural' Factors that Might Keep Wild Dog Numbers Low
  • Indirect Competition with other Large Carnivores
  • Direct Competition with other Large Carnivores
  • Predation by other Large Carnivores
• Human-induced Factors that Might Keep Wild Dog Numbers Low
  • Road Casualties
  • Direct Persecution
  • Snaring
• Diseases Affecting Wild Dogs
  • Viral Infections
  • Bacterial Infections
  • Protozoal Infections
  • Macroparasites
  • General patterns
• Conclusions

Background

In the previous chapter we showed that wild dogs have declined throughout Africa, principally as a result of habitat fragmentation and human persecution. However, a number of authors have remarked that, even in large, well-protected areas, wild dogs always live at very low densities (e.g. Mills & Biggs 1993; Schaller 1972). For example, lion densities are 3-20 times those of wild dogs, and spotted hyenas may outnumber wild dogs by factors varying from 8 to over a hundred (Table 4.1, Creel & Creel 1996). In this chapter, we review the factors thought to keep wild dogs' numbers low, and discuss how these problems may be compounded by habitat fragmentation. In the next chapter, we use demographic modelling to assess the extent to which each of these factors might threaten the long-term persistence of wild dog populations.

In the broadest terms, the size of a population will be defined by the rate at which individuals arrive in it - by birth and immigration - and the rate at which they leave it - by death and emigration. Local population decline will occur when recruitment is low and mortality or emigration rates are high. Therefore, to understand why wild dogs are so rare, and to assess whether their numbers are likely to decline still further, we need to understand the factors controlling recruitment, mortality and dispersal. Our efforts to do this are hampered, to some extent, by the availability of data. Relatively little is known about the factors which contribute to breeding success or failure in wild dogs. Similarly, data on the causes of dispersal are rather sketchy: since wild dogs may disperse over very large areas, it is often difficult to distinguish dispersal from death (Burrows et al. 1995; Ginsberg et al. 1995a). However, reasonably good data are available on mortality of both adults and juveniles - and juvenile mortality represents a very important component of recruitment. In Tables4.2 and 4.3, we have summarized the available data on causes of mortality in well-studied wild dog populations. These data form the basis of our discussion below. However, they should be interpreted with caution for two reasons. First, most of the study populations live inside or around national parks and game reserves and may not be representative of populations outside of protected areas. Second, to know the cause of an individual's death one must find the carcass, and this is likely to bias the results. At the extreme, one is more likely to find an adult killed in a road accident than a pup that dies of disease underground. Radio-telemetry greatly improves the probability of recovering a carcass, and therefore provides a less biased assessment of the causes of adult mortality. Indeed, Ginsberg et al. (1995a) found that such biases led to significant differences in the causes of mortality observed in collared and un-collared wild dogs.

In this chapter, we first outline the effect of 'natural' factors, such as competition with other large carnivores, likely to limit wild dog numbers. We then discuss the effect of human activities such as road accidents and persecution. The third and final section deals with the diseases that affect wild dogs. Since domestic dogs are the most important reservoir for canid diseases, it is often unclear whether disease represents a 'natural' or a human-induced threat to wild dogs.

'Natural' Factors that Might Keep Wild Dog Numbers Low

Indirect Competition with other Large Carnivores

The survival and reproductive success of a wild dog pack will depend, at least in part, upon its ability to secure prey. However, no wildlife communities are known to exist in which wild dogs are the only large predators: wild dogs coexist with other carnivores such as lions, spotted hyaenas, leopards and cheetahs. Wherever they have been studied, the spectrum of prey taken by wild dogs is very similar to that of other predators living in the same area (Creel & Creel 1996), raising the possibility that wild dogs might compete for prey with other carnivores. Specifically, other carnivores might reduce prey populations to such low levels that wild dogs are unable to locate and catch sufficient prey.

Where wild ungulates are abundant, such a scenario seems very unlikely. Ecological studies of wild dogs have suggested that their numbers are not limited by the availability of food (Ginsberg et al. 1995b; Mills & Biggs 1993). Wild dogs are efficient predators: they seldom seem to experience problems finding prey and have a high success rate when hunting (Creel & Creel 1995; Estes & Goddard 1967; Fanshawe & FitzGibbon 1993; Schaller 1972). Furthermore, wild dogs are crepuscular, while their possible competitors are either mainly nocturnal (hyaenas, lions, leopards) or diurnal (cheetahs, Mills & Biggs 1993). Competition might reduce wild dogs' hunting success in areas where ungulate prey are very scarce. However, it seems unlikely that indirect competition with other large predators has a substantial effect upon wild dogs in most areas where there are still resident populations.

Direct Competition with other Large Carnivores

Indirect competition probably has no substantial effect upon wild dog numbers. However, although they are efficient hunters, wild dogs do sometimes lose their kills to scavengers - indeed, a number of authors have suggested that one benefit of sociality for wild dogs is that group living allows for more effective defence of the kill (Kruuk 1975; Lamprecht 1978).

Although wild dogs occasionally lose their kills to lions, spotted hyaenas are much more important kleptoparasites (Creel & Creel 1996). For example, in the Serengeti National Park, Tanzania, hyaenas were present at 86% of wild dog kills and always fed from carcasses eventually (Fanshawe & FitzGibbon 1993). Conversely, in Serengeti wild dogs appropriated just 1% of hyaena kills (Kruuk 1972). Hyaenas seem to find it more difficult to locate wild dog kills in denser vegetation: in the Selous Game Reserve, Tanzania, hyaenas were present at only 18% of kills (Creel & Creel 1996). Nonetheless, wild dogs often go out of their way to mob hyaenas in Selous and elsewhere (Creel & Creel 1996).

Does this direct competition for kills have any detrimental effect upon wild dogs?Again, the answer varies among populations. In Serengeti, where hyaena density was high and wild dog kills highly visible, the presence of four or more hyaenas did reduce the time wild dogs were able to spend feeding from carcasses and, presumably, the amount that they ate (Fanshawe & FitzGibbon 1993). This effect was mitigated when more wild dogs were present: feeding time increased with the ratio of dogs to hyaenas. In contrast, in the thicker vegetation of Selous, where hyaena density was lower and relatively fewer hyaenas were attracted to wild dog kills, the presence of hyaenas had no effect on the time wild dogs spent feeding from each carcass. Hyaenas eventually fed from just 2% of wild dog kills in Selous, and wild dogs seemed to make no effort to avoid using areas frequented by hyaenas (Creel & Creel 1996).

Direct competition with hyaenas might depress wild dog numbers by reducing their feeding success - this might lead to both higher mortality and lower reproductive success, and, thus, to smaller populations. Fuller & Kat (1990) showed that wild dog packs have a relatively high food intake rate when they are feeding pups (average 4.1kg/dog/day with pups, compared with 1.6kg/dog/day without), and pointed out that one pack with a food intake rate similar to that of a pack without pups (2kg/dog/day) subsequently abandoned the litter that it was raising. Thus, it is possible that reduced feeding time as a result of harassment by spotted hyaenas might cause wild dogs to abandon their pups. Creel & Creel (1996) found a negative correlation between the population densities of wild dogs and spotted hyaenas across five study sites in eastern and southern Africa. Unfortunately, areas with high densities of hyaenas also have abundant lions, making it difficult to disentangle the effects of the two larger carnivores on wild dog numbers (see below).

Predation by other Large Carnivores

Although wild dogs are predators themselves, they are also the victims of predation. Twenty-two percent of adult mortality (16/74 deaths) and 42% (19/45) of juvenile mortality across study sites can be attributed to predation by other large carnivores (Tables 4.2 & 4.3). Of those animals killed, 75% (12/16) of adults and 89% (17/19) of pups were killed by lions. Predation by spotted hyaenas is less important: there are reports of just one adult and two juvenile wild dogs being killed by hyaenas (Tables 4.2 & 4.3), and the two pups were debilitated by anthrax (Creel et al. 1995). The relative importance of the two predators is reflected in wild dogs' response to them - wild dogs move away from the sound of lions roaring, but they mob hyaenas (Creel & Creel 1996).

Predation by lions is likely to have a marked effect upon wild dog populations, even though wild dogs form a negligible part of lions' diet. Field studies of community ecology indicate that predators are more likely to suppress the populations of prey that they kill only opportunistically (Erlinge et al. 1984). While predators will suffer themselves if they cause a reduction in the numbers of their favoured prey, they will compensate for the loss of less favoured prey by feeding upon other species. This may explain the finding that large predators can often limit the numbers of smaller predators, which form part of their diet (Polis & Holt 1992). African golden cats appear to be limited in part by leopard predation (Hart et al. 1996), swift foxes may be limited by coyotes (Carbyn et al. 1994), and predation by lions is the single most important cause of juvenile mortality in cheetahs (Laurenson 1994).

Does lion predation have any effect upon wild dog populations? Several lines of evidence suggest that it does. First, there is a correlation between the population densities of wild dogs and lions across four populations, with wild dog density highest where lions are scarce (Creel & Creel 1996). Unfortunately, areas with high densities of lions also have abundant hyaenas, making it difficult to disentangle the effects of the two larger carnivores on wild dog numbers (see above). Second, an attempt to reintroduce wild dogs to Etosha National Park, Namibia, failed when a pride of lions killed members of the introduced pack (Scheepers & Venzke 1995). Finally, a sudden crash in the population of lions in the Ngorongoro crater in the mid-1960s was followed by the appearance of wild dogs in the area. As the lion population recovered, wild dogs disappeared (Creel & Creel 1996).

Human-induced Factors that Might Keep Wild Dog Numbers Low

Road Casualties

The contribution of road traffic accidents to wild dog mortality varies between populations as a result, it seems, of the distribution and quality of roads. Where parks authorities keep speed limits low, and where roads are poor, very few wild dogs are hit by vehicles; only one of 23 adult wild dogs found dead in Kruger and Selous was killed by a vehicle (Table 4.2). However, road traffic accidents may be the single most important cause of adult mortality where wild dogs occupy areas with good roads used by fast-moving traffic. More than half the recorded adult mortality in Hwange National Park, Zimbabwe, is caused by accidents on the road between Bulawayo and Victoria Falls which runs along the northern edge of the Park (Table 4.2). In addition, three wild dogs were killed on a 20km stretch of the Tanzania-Zambia highway where it passes through Mikumi National Park, Tanzania, in a 15 month period (Drews 1995), and Tanzania National Parks records indicate that in one year 11 wild dogs were killed by vehicles passing through Mikumi (Creel & Creel 1993). In recent years eight wild dogs have been killed on the Lusaka-Mongu highway in Zambia, where it passes through Kafue National Park (K. Buk, pers. comm.).

Outside protected areas, road casualties are likely to cause relatively more wild dog deaths than inside them. For example, very few wild dogs use the area around Bulawayo, Zimbabwe, but two were killed within 30km of the city within a two year period (J.R.G., Unpublished data). Where roads are available, wild dogs use them to move and hunt. Indeed, road kills constitute an important source of information about the distribution of wild dogs living outside protected areas (See Chapter3).

Direct Persecution

Direct persecution by man has, perhaps, been the single most important cause of wild dogs' decline throughout Africa in the last century. Wild dogs were shot as vermin, even in national parks where, as Bere (1955) commented: ·...it was considered necessary, as it had often been elsewhere, to shoot wild dogs in order to give the antelope opportunity to develop their optimum numbers...º. Such shooting continued for many years; for example, wild dogs were shot by park staff until as recently as 1973 in Tanzania, 1975 in Zimbabwe, and 1979 in Niger (see Chapter3).

Although persecution of wild dogs is no longer national parks' policy, direct persecution by man remains an important cause of mortality even in populations inhabiting protected areas: Table4.2 shows that shooting and poisoning accounted for the deaths of 28/105 (27%) adult wild dogs across five areas - and four of these areas are at least partially protected. Local people are also known to poison wild dogs in the Maasai steppe, in Tanzania (Fitzjohn 1995).

Wild dogs are persecuted where they are perceived as a pest which kills livestock, or competes with people for wild ungulates in hunting areas. For example, an unconfirmed report suggests that over 50 wild dogs were shot on a hunting concession outside Hwange National Park between 1987 and 1991. Such persecution represents an important cause of mortality, even for dogs which spend much of their time inside the Park.

The available evidence suggests that wild dogs' reputation as voracious stock-killers is rarely justified (Bowler 1991). Livestock are taken occasionally but, where wild prey are available, losses to farmers seem to be small, especially for larger livestock. The only systematic study of this problem found that, over a two-year period, wild dogs took just 26 cattle from a herd of 3,132 in the Nyamandhlovu region of Zimbabwe, and all of these were calves and weaners rather than adults (Rasmussen 1996). Losses to wild dogs accounted for just 1.8% of the combined financial cost of all livestock losses. However, losses of small stock may be dramatic: one pack of wild dogs killed 70 ewes and 67 lambs on a single ranch in Laikipia in 1996 (M. Dyer pers. comm.). As for other canids (Ginsberg & Macdonald 1990), levels of stock loss to wild dogs may be low overall, but a few farms tend to suffer disproportionately and local losses may be severe.

Nevertheless, if wild prey are available wild dogs usually ignore livestock (Fuller & Kat 1990) - indeed, on one occasion wild dogs in Nyamandhlovu passed through a calf paddock to chase a kudu in the adjacent paddock (Rasmussen 1996). Despite these low stock losses, farmers in this area of Zimbabwe wanted the wild dogs killed. Thus, persecution remains a serious problem for wild dogs living in unprotected areas. Farmers are known to shoot wild dogs in most places where they occur outside protected areas. In countries where wild dogs survive mostly outside protected areas, such as Namibia, Kenya and Ethiopia, such persecution must represent a very serious threat to their long term survival. Since packs using parks and reserves may also make frequent and extensive forays into unprotected areas, they are also vulnerable to persecution.

Snaring

Snares cause a significant proportion of wild dog mortality, even for populations living inside protected areas: 10/105 (10%) of adult deaths were caused by snares (Table4.2). Snares are less of a problem for pups, causing the deaths of only 2/45 pups (4%).

In most places, snares are not set to catch wild dogs: they are caught accidentally in snares set for ungulates. Thus, wild dog mortality is an incidental effect of subsistence hunting outside protected areas, and poaching inside them. Wild dogs living in parks and reserves often encounter snare lines as they move out into unprotected areas (S.Creel pers. comm.) - a similar phenomenon is common in spotted hyaenas (Hofer et al. 1993).

In some areas of Zimbabwe parts of wild dogs are used for ritual and medicinal purposes - thus snares are set specifically to catch wild dogs (J.R.G., Unpublished data). Such snares may cause very high mortality within individual packs.

Diseases Affecting Wild Dogs

The threat that disease poses to endangered species has been recognized more and more in recent years (Dobson & Hudson 1986; Karesh & Cook 1995). For example, canine distemper brought the black-footed ferret to the brink of extinction (Williams et al. 1988), and a similar disease has been implicated in the extinction of the thylacine (Guiler 1961). Might disease, then, pose a threat to the remaining wild dog populations?

Many authors have noted wild dogs' susceptibility to disease, and suggested that this might help to explain their low densities (e.g. Bere 1955; Schaller 1972). This makes it surprising that Tables 4.2 & 4.3 show little evidence of disease-induced mortality: only 8 of 74 adults (11%), and 5 of 45 pups (11%) are believed to have died from disease across study sites. One reason for this apparent paradox is that the mortality from disease is mostly episodic in wild dogs: numbers might remain stable for several years, but then a single epizootic may cause sudden dramatic decline or even local extinction. The data presented in Tables 4.2 & 4.3 come from stable populations unaffected by epizootics at the time of study. Other studies (for which systematic mortality data are not available) show a different picture. Rabies caused the death of 21 of the 23 wild dogs in the Aitong pack outside the Masai Mara National Reserve, Kenya, leading to the extinction of the pack in a period of just 44 days in 1989 (Kat et al. 1995). By June 1991, the whole wild dog study population of the Masai Mara and the contiguous Serengeti National Park, Tanzania - a total of eight packs - had disappeared, with disease suspected or confirmed in each case. Disease was therefore believed to have caused the extinction of the wild dog study population in the Serengeti ecosystem (see Appendix 1). Disease also seems to have caused local population decline in other areas. For example, sightings of wild dogs declined dramatically after an outbreak of anthrax in ungulates in the Luangwa Valley, Zambia, which is also known to have killed wild dogs (Turnbull et al. 1991), and population declines of wild dogs in north-west Zimbabwe in the early 1980s coincided with an epidemic of rabies in jackals (Childes 1988; Kennedy 1988).

In the following sections, we detail the pathogens which are known to infect free-ranging populations of wild dogs. In Table 4.4, we present data on the prevalence of infection with these pathogens where such data are available. It should be borne in mind that many of these data depend upon serology; that is, the data show which animals have antibodies to the various pathogens or to the toxins they secrete, but give no information about how or when the animals were exposed to the pathogens. The proportion of seropositive animals within a population is affected by a number of factors. A high seroprevalence could indicate that most animals become infected early in life, but that the resulting disease is mild and most animals recover and become immune. Alternatively, the same seroprevalence could indicate that the population has recently experienced an epidemic of a highly virulent disease, and that only those that survived infection (and are thus seropositive) remain in the population. The pattern of seroprevalence in different age classes can help to distinguish between these alternatives (Thrusfield 1986). However, the sample sizes for wild dogs are rarely large enough to allow assessment of such patterns. In the absence of these data, we have inferred the likely impact of each pathogen from observations of wild dogs in the field and in captivity, and from the effect of each disease upon domestic dogs (Table 4.5).

We have designated the pathogens known to cause substantial mortality in wild dogs with the symbol . The effects of the various pathogens are also summarized in Table 4.5. A number of patterns emerge from this survey, which we discuss at the end of the section. The possible impacts of some of these pathogens on population persistence are investigated in the following chapter.

Viral Infections

Rabies Virus

Rabies is a rhabdovirus which may infect all mammals. In North America and Europe, populations of wild carnivores such as racoons and red foxes represent the major reservoir for the virus, but in Africa, as well as Asia and South America, poorly supervised domestic dogs are the principal host (Baer & Wandeler 1987). Rabies represents a major threat to endangered canids: one epidemic halved the population of Ethiopian wolves in the Bale Mountains National Park, Ethiopia (Sillero-Zubiri et al. 1996), while another threatened the Blanford's fox in Israel (Macdonald 1996).

Rabies is known to cause high mortality in wild dogs. In 1989, a well-studied pack living at Aitong, outside the Masai Mara National Reserve, Kenya, was decimated by rabies (Kat et al. 1995). The following year, at least one wild dog died of rabies in the adjoining Serengeti National Park, Tanzania (Gascoyne et al. 1993). Wild dog packs under study in the Serengeti ecosystem disappeared in 1991, and, although the ultimate cause is not certain, rabies is the most likely culprit (Burrows 1992). The circumstances surrounding the Serengeti extinction are discussed in detail in Appendix 1. Rabies is also known to have killed wild dogs in the Central African Republic (A.K.Turkalo pers. comm.) and in Namibia (Scheepers & Venzke 1995), and is believed to have killed dogs in Zimbabwe (C.M. Foggin, cited in Kat et al. 1995) and Zambia (K.Buk pers. comm.).

Rabies virus is transmitted principally by biting. In the Aitong pack, infected animals joined in with group activities such as greetings and cooperative hunting, but were often attacked by other group members (Kat et al. 1995). This led to biting and, presumably, transmission of the virus. Infected animals became disoriented and lost their appetites, but chewed and consumed non-food items. They became ataxic and progressively paralysed (Kat et al. 1995). These symptoms are similar to those of 'dumb' rabies in domestic dogs (Baer & Wandeler 1987).

The few data available on rabies dynamics in wild dogs suggest that the infection would be unlikely to persist in their populations. The disease spread rapidly through the Aitong pack: the time from the first suspected infection of a single pack member to the death of the last of the 21 dogs that died was less than two months (Kat et al. 1995). Since transmission of the virus between pack members is rapid, the incubation period is short, and mortality seems very high, the virus would probably cause its own local extinction before it could be transmitted to another pack (Kat et al. 1995; Mills 1993). Rather than persisting in wild dogs, rabies is probably maintained in the populations of other hosts, which act as a reservoir from which infection occasionally spills over into wild dogs. Rabies is endemic in the domestic dog populations of some areas surrounding the Serengeti ecosystem (Cleaveland & Dye 1995), and the virus which decimated the Aitong pack was genetically indistinguishable from one isolated from local domestic dogs (Kat et al. 1995). Thus, in this case domestic dogs appear to have been the reservoir host for rabies. However, in southern Africa wild canids, such as jackals and foxes, may be more important in maintaining the infection (Nel 1993).

Canine Distemper Virus

Canine distemper virus is a morbillivirus related to rinderpest, human measles, and phocine distemper, which is transmitted by inhalation of airborne viral particles (Appel 1987c). The virus attacks most terrestrial carnivores, and in the past it has led to dramatic declines in populations of black-footed ferrets (Williams et al. 1988) and lions (Roelke-Parker et al. 1996). Wild dogs' susceptibility to canine distemper virus has been demonstrated on several occasions when vaccination of captive animals with live attenuated vaccines has been followed by distemper-like disease and death (Durchfeld et al. 1990; McCormick 1983; van Heerden et al. 1989).

There is only one confirmed case of free ranging wild dogs' dying of canine distemper - ten died in northern Botswana in 1994 (Alexander et al. 1996). However, circumstantial evidence suggests that distemper has caused the deaths of many wild dogs in the past. Schaller (1972) described how members of one pack in Serengeti contracted a disease which resembled ·...a typical picture of the gastrointestinal form of distemper...º. However, neither canine distemper virus nor antibodies were identified, so this diagnosis remains unconfirmed. Reich 1981 (cited in van Heerden et al. 1995) also reported nervous symptoms of a disease resembling canine distemper in wild dogs in Kruger although, again, the diagnosis was not confirmed. A wild dog showing symptoms of canine distemper was seen in Hluhluwe-Umfolozi Park in 1995 (J.van Heerden pers. comm.). Finally, the extinction of wild dogs in the Serengeti/Masai Mara area in 1990-1 has been attributed to an epidemic of canine distemper (Alexander & Appel 1994; Macdonald et al. 1992), although other authors have contested this (Burrows et al. 1995). This possibility is discussed in detail in Appendix 1.

Serological surveys indicate, however, that canine distemper infection is not always fatal for wild dogs. High seroprevalences have been recorded recently in Hluhluwe-Umfolozi Park, in Northern Botswana, in the Selous Game Reserve, and in Tsumkwe District, Namibia (J. van Heerden & J.W. McNutt, pers. comm., Creel et al. in prep.; Laurenson et al. in prep.). indicating that some wild dogs had contacted the virus and survived. The mortality caused by canine distemper infection is not clear. No signs of distemper-related mortality or sickness have been recorded in Selous, despite intensive monitoring (Creel et al. in prep.). Possible evidence of disease has been seen in Hluhluwe, however (J.van Heerden pers. comm.), and at least one pack in Northern Botswana was decimated by canine distemper (Alexander et al. 1996).

It has been suggested that, as for rabies, domestic dogs may act as a reservoir host for canine distemper. Indeed, in areas where wild dogs are known or suspected to have been infected with canine distemper, local domestic dog populations show high seroprevalence for canine distemper virus (Table 4.6). However, wild dogs also show a high prevalence of antibodies to canine distemper in Selous, even though domestic dogs (and other wild canids) are very rare. The nearest concentration of domestic dogs is in Morogoro, some 70km from Selous, where domestic dogs have experienced canine distemper (S.R. Creel pers. comm.). Thus, it appears that canine distemper may be persisting in Selous without recourse to a domestic dog reservoir. If the infection is persisting in the wild dogs themselves, it is possible that the viral strain has a relatively low pathogenicity for wild dogs (Creel et al. in prep.). Alternatively, some other wild carnivore might be acting as a reservoir. More research is needed to reveal the impact of canine distemper infection on free-ranging wild dog populations.

Canine Parvovirus

Canine parvovirus is a virus that replicates only in canids. It appeared, apparently by mutation, in the late 1970s and spread rapidly to domestic dogs world-wide (Appel & Parrish 1987). Antibodies to the virus have been found in wild dogs in Serengeti and Selous (M.K.Laurenson, pers. comm., Creel et al. in prep.) and in the Masai Mara region (Alexander et al. 1993), but not in Kruger (van Heerden et al. 1995) or Tsumkwe District, Namibia (Laurenson et al. in prep.).

In domestic dogs, parvovirus replicates principally in the dividing cells of the intestinal epithelium, and the resulting enteritis may be an important cause of mortality in puppies. Infected dogs excrete viral particles in their faeces, and these viruses may persist in the environment for relatively long periods of time (Appel & Parrish 1987).

It is not known whether parvovirus persists in wild dog populations or whether, like rabies, it 'spills over' from domestic dogs. Wild dog populations in the Masai Mara and Tsumkwe had lower seroprevalences than sympatric domestic dogs (Masai Mara: 7% of wild dogs (n=15) and 25% of domestic dogs (n=181) seropositive, Alexander et al. (1993); Tsumkwe: 0% of wild dogs (n=6) and 47% of domestic dogs (n=70) seropositive, Laurenson et al. (in prep.). However, in Selous the infection appears to persist in the absence of domestic dogs (Creel et al. in prep.).

The impact of parvovirus on wild dog populations remains unknown. Long-term studies of grey wolves show that, while parvovirus infection is an important cause of juvenile mortality, the effect on recruitment is not sufficient to cause a population decline (Mech & Goyal 1995). The virus is, however, believed to have hindered the recovery of some wolf populations (Mech & Goyal 1995). Thus, parvovirus might help to keep wild dog populations small, especially in fragmented populations that have frequent contact with domestic dogs.

Canine Adenovirus (Infectious Canine Hepatitis)

Infectious canine hepatitis is a disease of domestic dogs and other canids caused by Type1 canine adenovirus, a DNA virus. Antibodies to canine adenovirus have been found in wild dogs in Kruger (van Heerden et al. 1995), as well as Serengeti and the Masai Mara (M.K. Laurenson, pers. comm.; K. Alexander, Unpublished data).

A high proportion of wild dogs sampled in Kruger carried antibodies to the virus. Similar patterns of seroprevalence come from infected populations of domestic dogs: most animals become infected early in life and acquire immunity without showing signs of disease (Appel 1987a). However, mortality may be very high in young puppies. Thus, it seems unlikely that canine adenovirus has much effect upon adult wild dogs, but it might be a cause of juvenile mortality.

Canine Coronavirus

Canine coronavirus is a virus that replicates only in canids. Antibodies to the virus have been found in wild dogs from Kruger (van Heerden et al. 1995), and the Masai Mara (K. Alexander, Unpublished data). On its own, coronavirus causes a mild gastroenteritis in domestic dogs; however, mixed infections with parvovirus are common and may be fatal (Appel 1987b). Like parvovirus, coronavirus particles are excreted in the faeces and contact with infected faeces represents the most important route of transmission. In domestic dogs, disease occurs mainly in puppies, while infected adults rarely show signs of ill health. Although the effect of coronavirus infection on wild dogs remains unknown, it might be expected to follow a similar pattern.

Canine Herpesvirus

Canine herpesvirus is a DNA virus which replicates only in canids, and may cause high mortality in newborn puppies (Appel 1987d). Adult domestic dogs rarely show clinical signs of disease, although in infected populations most are seropositive (Appel 1987d). Antibodies to canine herpesvirus have been found in wild dogs in the Masai Mara (K. Alexander, Unpublished data). Any effect of the virus on wild dog populations remains unknown although, by extrapolation from domestic dogs, it seems likely that it affects juvenile rather than adult mortality.

Canine Para-influenza Virus

Canine para-influenza virus is a virus affecting domestic dogs, where it is one of the main causes of 'kennel cough' (Appel & Binn 1987). Antibodies to this virus - or possibly the closely related Simian Virus 5 - have been recorded from wild dogs in Kruger (van Heerden et al. 1995). In domestic dogs, infection with para-influenza virus alone leads to mild respiratory disease or, more usually, causes no clinical signs. However, under natural conditions infection is often accompanied by secondary infections by other viruses and bacteria (Appel & Binn 1987). The effect of the virus on wild dogs remains unknown, but is likely to be mild.

Reovirus

Three types of reovirus have been isolated from domestic dogs, but none appears to lead to a specific disease (Appel 1987f). Antibodies to reovirus are commonly found in domestic dogs, and have been recorded in wild dogs in Kruger (van Heerden et al. 1995). Although reovirus alone seems not to cause disease, dual infection with canine parvovirus and canine distemper does occur in domestic dogs. It is possible that reovirus has an immunosuppressive effect (Appel 1987f). It seems unlikely, though, that infection with reovirus has any marked effect on wild dog populations.

Rotavirus

Rotavirus, like reovirus, appears not to cause disease in domestic dogs (Appel 1987e). The finding of antibodies in wild dogs from Kruger is the first record of rotavirus infection in a wildlife population (van Heerden et al. 1995). It seems unlikely that this virus has any marked effect upon wild dog populations.

African Horse Sickness Virus

African Horse Sickness is an important disease of horses and other equids, including zebras. However, other species, including domestic dogs, may also carry the virus. The first survey of wild carnivores revealed antibodies in four populations of wild dogs, as well as sympatric lions, hyaenas, cheetahs and jackals (Alexander et al. 1995). African Horse Sickness is caused by an arbovirus which is transmitted between equids by Culicoides midges and mosquitoes. However, domestic dogs may contract the virus by eating infected meat (Losos 1986) and this seems the most likely route of infection for wild carnivores - seroprevalences are high in wild carnivores that prey on zebras (hyaenas, lions, wild dogs), but much lower in sympatric populations of domestic dogs (Alexander et al. 1995).

It is not known whether infection with African Horse Sickness virus has any effect on wild dogs, but it can cause illness and mortality in domestic dogs. It seems unlikely, however, that this virus has any marked effect upon wild dog populations.

Bluetongue Virus

Bluetongue is primarily a disease of sheep, in which it can cause dramatic economic losses (Losos 1986). The bluetongue virus also affects several wild ruminant species, and antibodies to the virus were recently isolated from wild dogs for the first time (Alexander et al. 1994). Antibodies were present in all four wild dog populations that were surveyed. Bluetongue is caused by an arbovirus closely related to the one that causes African horse sickness. Like African horse sickness, bluetongue is usually transmitted by Culicoides midges, but eating infected meat is probably the most important route of infection for predators. The virus is fairly resilient and remains viable even in decomposed blood (Losos 1986).

It is not known whether infection with bluetongue virus has any adverse effects on wild dogs, but it has caused abortion in domestic dogs (Alexander et al. 1994). It seems unlikely, however, that this virus has any marked effect upon wild dog populations.

Bacterial Infections

Bacillus anthracis (Anthrax)

Anthrax is an extremely important bacterial disease that affects most mammals. Although a serological survey of a small sample of wild dogs in Kruger showed no evidence of exposure to the disease (van Heerden et al. 1995), anthrax is known to have killed wild dogs in Kruger, as well as in Selous (Creel et al. 1995), and in South Luangwa National Park, Zambia (Turnbull et al. 1991).

The spores of Bacillus anthracis may survive in the soil for years, so the pathogen can persist in an area even in the absence of a reservoir host (Turnbull 1990). Animals in the final stages of anthrax haemorrhage from the nostrils, mouth and anus, and bacteria in the blood sporulate on contact with the air. As a result, ungulates usually become infected by contact with bacterial spores in the soil or water (Turnbull 1990). However, carnivores become infected by eating the flesh of infected animals. Some carnivores appear highly resistant to the disease: for example, during a serious anthrax epidemic in Etosha National Park, Namibia, lions, spotted and brown hyaenas, and black-backed jackals all fed from the carcasses of animals which had died from anthrax, but showed no signs of the disease themselves (Ebedes 1976). Similarly, during an epidemic in the Luangwa valley in 1987, one area of just 80km² yielded the carcasses of 101 hippos, 60 buffalo and 20 elephants, along with puku, kudu and other ungulates - but only one spotted hyaena and two leopards (Turnbull et al. 1991).

Wild dogs' resistance to anthrax seems to vary. The Luangwa epidemic was accompanied by a marked decrease in the frequency of sightings of wild dogs throughout the Park. Five carcasses of wild dogs were found, and anthrax was confirmed in four of them (Turnbull et al. 1991). It seems likely, therefore, that the population decline can be directly attributed to anthrax. However, anthrax does not always have such marked effects upon wild dogs. Anthrax epidemics occurred in Kruger in 1990, 1991 and 1993, but the wild dog population in the area increased during this period, and only 3 of 1538 anthrax-positive carcasses were wild dogs (M.G.L. Mills pers. comm., de Vos & Bryden 1996).

Anthrax has also been reported from a wild dog pack in Selous (Creel et al. 1995). Three adults and eight pups, from a group of 18 adults and 24 pups, showed signs of disease. All of the adults recovered, but four of the pups died. Thus, wild dogs can recover from anthrax - indeed, animals which had shown signs of disease were no more likely to die in the six months following the outbreak than were apparently uninfected animals. The outbreak had no effect on the pack's movement patterns or hunting success. Furthermore, there was no transmission of the infection between pack members, although apparently healthy animals licked saliva and ocular discharge from the faces of sick pups. However, this outbreak did not take place during an anthrax epidemic in the ungulate prey base, and was probably caused by some members of one pack killing and consuming a single animal that harboured enough bacilli to transmit the disease (Creel et al. 1995). Under epidemic conditions wild dogs would be exposed to prey infected with anthrax repeatedly, and it is possible that a greater proportion of wild dogs in each pack might have been affected. Thus, anthrax may sometimes have a dramatic effect upon wild dog populations, but this is certainly not always the case.

Ehrlichia canis (Ehrlichiosis)

Ehrlichiosis is a disease of domestic dogs, caused by the rickettsial bacterium Ehrlichia canis and transmitted by the brown dog tick, Rhipicephalus sanguineus. This disease was believed to have contributed to the decline of wild dogs in Kruger in the 1920s and 1930s (Stevenson-Hamilton 1939). At that time, many domestic dogs living in the park died of ·...a disease against which the usual treatment for biliary fever and distemper seemed to be of no avail...º (Neitz & Thomas 1938). Blood slides taken from two domestic dogs that contracted the disease contained Ehrlichia canis and also, subsequently, Babesia canis (see below). Local people reported having seen wild dogs showing the same symptoms, but ehrlichiosis was not confirmed (Neitz & Thomas 1938). van Heerden (1979) showed experimentally that wild dogs can contract ehrlichiosis, although the disease was less severe in wild dogs than in domestic dogs. Surveys of wild dogs in Kruger and the Masai Mara have found no evidence of exposure to Ehrlichia canis, although a few domestic dogs in the Masai Mara were seropositive (Alexander et al. 1993; van Heerden et al. 1995). Thus, any effect of ehrlichiosis on free-ranging wild dog populations remains obscure.

Rickettsia conorii/ africae (Spotted Fever)

Spotted fevers are a group of tick-borne diseases caused by some of the bacteria in the genus Rickettsia. A high proportion of wild dogs in Kruger show evidence of having been exposed to infection, although the two species occurring in Southern Africa, R.conorii and R.africae cannot be distinguished by serological means (van Heerden et al. 1995). Domestic dogs and other domestic mammals may become infected, but they show no clinical signs of disease (Marmion 1990). It seems unlikely, therefore, that spotted fever rickettsiae have any marked effect upon wild dog populations (van Heerden et al. 1995).

Coxiella burnetti (Q Fever)

Q fever is a disease of man, caused by Coxiella burnetti, an intracellular bacterium related to Rickettsia (Losos 1986). Many other wild and domestic mammals and birds may sustain infection, and antibodies were found in wild dogs from Kruger in 1990-3 (van Heerden et al. 1995). Mammals other than man usually show no clinical symptoms, although infection may occasionally cause abortion in sheep and goats (Losos 1986). It seems unlikely, therefore, that Coxiella infection has any substantial effect on wild dog populations.

Brucella abortus (Brucellosis)

Brucellosis is a commercially important disease which causes abortion and infertility in cattle. One of three wild dogs shot in Serengeti in 1965-7 showed evidence of previous infection with Brucella abortus, the bacillus which causes brucellosis (Sachs et al. 1968). This animal would almost certainly have contracted the infection by eating infected meat: the disease was widespread in zebra, wildebeest and other prey species at the time. Brucella canis causes abortion in domestic dogs, but the effect of Brucella abortus on wild dogs is not known. It seems unlikely, however, that this infection has any significant impact on wild dog populations.

Protozoal Infections

Toxoplasma gondii

Toxoplasma gondii is a sporozoan parasite which primarily affects cats, although other mammals can become infected. All wild dogs sampled in Kruger were seropositive for Toxoplasma (van Heerden et al. 1995). Four pups necropsied in Kruger were found to have died from an infection of either Toxoplasma or the closely related Neospora; 16 other pups from the same den disappeared at the same time (M.G.L. Mills & J. van Heerden, pers. comm.), although adult group members were not affected. Thus, Toxoplasma may cause some juvenile mortality, but seems not to affect adult wild dogs.

Neospora caninum

Neospora caninum is a sporozoan parasite related to Toxoplasma, which was first discovered in 1978. In domestic dogs it may cause paralysis in pups, and also abortion (Ruehlmann et al. 1995). Infection has not been confirmed in wild dogs, but four pups necropsied in Kruger were found to have died from an infection of either Neospora or the closely related Toxoplasma; 16 other pups from the same den disappeared at the same time. Thus, Neospora might cause some mortality in wild dog pups.

Babesia

Babesiosis is a tick-borne disease caused by intraerythrocytic protozoa of the genus Babesia. The parasite affects many species of wild and domestic mammals (Losos 1986), and has been recorded from wild dogs in Kruger (van Heerden et al. 1995), and probably also Serengeti (Peirce et al. 1995). Captive wild dogs usually carry the parasite without showing signs of disease (van Heerden 1980), although one pup died in captivity as a result of acute babesiosis (Colly & Nesbit 1992). Thus, Babesia infection might cause disease in wild populations, but it seems unlikely that it has any substantial effect on wild dog numbers.

Hepatozoon

Hepatozoon is a genus of apicomplexan protozoa than infects a wide range of vertebrates. Infestation may be severe in domestic dogs suffering from other infectious diseases such as ehrlichiosis. The parasite has been recorded in wild dogs in Kruger (van Heerden et al. 1995) and Serengeti (Peirce et al. 1995). It is not known whether Hepatozoon infection has any adverse effects on wild dogs, but domestic dogs infected with the parasite usually show no clinical signs of disease (van Heerden et al. 1995). However, the parasite infects the white blood cells and presumably causes some impairment of the immune system. Nevertheless, it seems unlikely that Hepatozoon has any substantial effect upon wild dog populations.

Macroparasites

As well as the viral, bacterial and protozoal infections discussed above, wild dogs are also hosts for a number of macroparasites. The hookworm Ancylostoma caninum has been found in wild dogs from Kruger, the Masai Mara, Moremi and Hwange (Spangenberg & Ginsberg Unpublished data, van Heerden et al. 1994). This nematode has caused illness in captive wild dog pups. In Serengeti and Hwange, wild dogs often 'anal dragged' - a typical behaviour of domestic dogs infected with intestinal parasites. One animal which often showed this behaviour in Serengeti also appeared bloated, and lacked stamina when hunting (J.R.Mal-colm pers. comm.). Thus, infection with macroparasites might be a contributing factor to mortality of young or malnourished wild dogs. However, it seems unlikely that they have any substantial effect upon wild dog populations.

General Patterns

Two patterns emerge from this survey of wild dog diseases, which point to the need for concern and, in some cases, more research.

First, many of the diseases affecting wild dogs are likely to have been contracted from sympatric domestic dogs. Domestic dogs are believed to act as reservoir hosts, from which diseases 'spill over' into wild dog populations: since wild dogs live at such low densities, it is unlikely that pathogens causing significant mortality could persist in their populations in the absence of such a reservoir. This possibility leads to further concern. Epidemiological models of diseases infecting more than one host within a community usually predict the extinction of species which are more affected by transmission from other species than by transmission from members of their own species (Begon & Bowers 1995). More research is needed in this direction if appropriate strategies for disease control are to be formulated.

Second, most of our knowledge of wild dog diseases is based upon serology, which shows only whether an animal has been exposed to a particular pathogen in the past. Even if an animal is found to be seropositive, the timing of the infection and its effects upon the host remain unknown. Furthermore, animals which die from exposure to the same infection do not, by their very nature, show up in serological surveys. As a result, the effects of many pathogen species on the health of individual wild dogs and the characteristics of wild dog populations remain unknown. For example, canine distemper appears highly pathogenic to wild dogs held in captivity, and yet some free-ranging populations show a high seroprevalence, indicating that animals have survived exposure to the disease. Without knowing the mortality caused by such a disease, it is difficult to assess its likely impact upon wild dog populations.

Similarly, wild dog populations show high seroprevalences for a number of viral infections thought likely to contribute to pup mortality. However, it is difficult to assess their impact since young pups usually remain in the den, making it difficult (and, in all probability, unethical) to sample them.

Conclusions

This discussion has revealed a number of potential threats to the remaining populations of African wild dogs. Perhaps the most important conclusion is that human presence poses a serious threat to wild dogs, even in the largest and best-protected areas: 61% of recorded adult mortality is caused directly by human activity (Table4.2). Wild dogs using protected areas may range outside the borders and into areas used by people. Here they encounter high-speed vehicles, guns, snares and poisons, as well as domestic dogs which may represent reservoirs of potentially lethal diseases.

The important rôle played by human-induced mortality has two long-term implications. First, it makes it likely that, outside protected areas, wild dogs may well be unable to co-exist with the rising human population unless better protection and local education programmes are implemented. This will be a serious problem for wild dog populations in areas such as Ethiopia and Namibia, where most populations occur outside protected areas. Second, wild dogs' ranging behaviour leads to a very substantial 'edge effect', even in large reserves. Simple geometry dictates that a reserve of 5,000km² can contain no point less than 40km from its borders - a distance well within the range of distances travelled by wild dogs in their usual behaviour. Thus, a reserve of this size (fairly large by most standards) would be, from a wild dog's perspective, all edge. As human populations rise around reserve borders, the risks to wild dogs venturing outside are also likely to increase. Under these conditions, only the very largest reserves will be able to provide any level of protection for wild dogs.

Even in large, well-protected reserves, wild dogs live at very low population densities. It seems likely that predation by lions, and, perhaps, competition with hyaenas, contribute to keeping wild dog numbers below the level that their prey base might support. Even within large parks such as Tsavo West in Kenya, wild dogs appear to select certain habitat types in which to live. Such low population density brings its own problems. The largest areas contain only relatively small wild dog populations; for example the Kruger National Park and surrounding reserves, with a combined area of 26,000km² (about the size of Israel), contain just 375 wild dogs (Maddock & Mills 1994). Most reserves, and probably most wild dog populations, are smaller: for example Niokolo-Koba National Park, at 9,000km², contains 50-100 wild dogs (C. Sillero-Zubiri, pers. comm.). Such small populations are vulnerable to extinction (Soulé 1987). 'Catastrophic' events such as outbreaks of epidemic disease may drive them extinct when larger populations would recover - such an event seems to have led to the extinction of the small wild dog population in Serengeti (Appendix1). Such problems of small population size will be exacerbated if, as seems likely, small populations occur in small reserves or habitat patches. As discussed above, animals inhabiting such areas suffer a strong 'edge effect'. Thus, small populations might be expected to suffer disproportionately high mortality as a result of their contact with humans and human activity.

Low population density may also cause problems related to disease transmission. Many diseases of domestic dogs appear to 'spill over' into wild dog populations, which probably occur at densities too low to allow the infection to persist. General models of similar systems predict the extinction of the host into which the disease 'spills over' - in this case wild dogs (Begon & Bowers 1995). Similar models designed specifically for wild dogs are needed to examine this problem in more detail.

One further problem related to disease is that wild dogs' social organization might hamper selection for disease resistance. In most animals, naturally resistant animals that survive disease outbreaks will experience reduced competition and high reproductive success after the epidemic. In this way, genes for resistance will spread in the population. However, survivors of local epidemics in wild dogs populations may rarely be able to pass on their genes for disease resistance. If only one or two pack members survive (as, for example, in the rabies outbreak in the Aitong pack, Kat et al. 1995), they will have to join or form a new pack if they are to have any hope of breeding. Such dispersing animals are believed to suffer high mortality in some areas (Ginsberg et al. 1995a), making it unlikely that pack remnants will survive long after the decimation of their packs. Thus, natural selection for resistance against epidemic diseases such as rabies may be weak in wild dogs.

To conclude, many factors, both natural and human-induced, conspire to keep wild dog numbers low. It seems likely that these threats will be compounded by habitat fragmentation, which will divide wild dogs into smaller populations each at disproportionate risk from human activities. In the next chapter, we use demographic modelling to investigate the likely impact of each of these factors on population persistence.

Return to Wild Dog Action Plan Table of Contents, AWD Species Account, CSG Publications or CSG Home Page.

© 1997 International Union for the Conservation of Nature and Natural Resources.