Chapter 6
Measures for the Conservation and Management of Free-ranging Wild Dog Populations

Rosie Woodroffe & Joshua R. Ginsberg

Previous chapters show that fragmentation, persecution, disease and road accidents represent serious threats to small wild dog populations, but that these risks diminish in larger populations. This suggests three paradigms for the management of Africa's remaining wild dog populations:

1. Maintaining large (>10,000km²) contiguous tracts of land set aside for wildlife represents the single most important strategy for wild dog conservation. Such areas are large enough to support viable wild dog populations, and contain core areas where wild dogs are fully protected from human activities. Measures that would benefit wild dogs include:

• maintaining the integrity of large protected areas
• establishing cross-border parks
• linking reserves by corridors
• establishing networks of smaller protected areas linked by privately, publicly, or communally held land managed for wildlife

Inside such wildlife areas, wild dogs would be protected by routine reserve management including:

• control of poaching to maintain their prey base
• severe restrictions on building high-speed roads in wildlife areas
• zero tolerance of domestic dogs - strays must be shot on sight.

(2) Integrated carnivore management programmes should be established to resolve conflicts between people and wild dogs where they coexist. Such programmes could involve:

• zoning of lands to define areas where predators will, and will not, be tolerated
• assessment of predator impact on livestock and wild prey species
• local conservation organizations working with farmers to minimize livestock losses through better husbandry practice
• compensation programmes for stock that are killed
• control, and perhaps vaccination, of domestic dog populations
• a ban on sport hunting of wild dogs

(3) Establishing tiny populations in small, fenced reserves may be the only way to conserve wild dogs in highly fragmented landscapes. Persistence would be improved by managing several such populations together as a metapopulation, periodically translocating animals between reserves. Such intensive management would be expensive and, while valuable for increasing the number of wild dogs in a local area or country, provides no substitute for protection of free-ranging populations.

• Protection of Wild Dog Habitat
  • National Parks and Reserves
  • Other Wildlife Areas
• Controlling Human-Induced Mortality
  • Persecution
  • Snaring
  • Road Traffic Accidents
• Managing the Threat of Disease
  • Minimizing Contact between Wild Dogs and Disease Reservoirs
  • Eradicating Diseases from their Reservoir Hosts
  • Vaccinating Wild Dogs Themselves
  • Choosing the Best Strategy for Disease Control
• Conclusions

Background

In previous chapters, we have described how wild dogs have been extirpated across much of Africa, and discussed the factors which threaten populations of various sizes. In this chapter, we use this information to propose measures for the conservation and management of the wild dogs that remain in Africa. In the next chapter, we discuss the possibilities for re-establishing populations by the reintroduction of wild dogs to areas where they have been extirpated.

Protection of wild dog habitat

Wild dogs only persist in countries with low human population density (Chapter3). Some wild dog populations do coexist with people - but such coexistence is only likely to be stable under certain circumstances:

1. The density of wild ungulate prey must remain high.
2. The density of domestic dogs must remain low. High density domestic dog populations can act as reservoirs for diseases that threaten wild dogs.

Such conditions mainly occur where human settlement has been curbed, either because the area has been set aside for wildlife, or through some external factor (e.g. tsetse flies, Rogers & Randolph 1988). Conservation of wild dogs therefore depends upon the long-term persistence of large areas where human population density remains low.

National Parks and Reserves

Maintaining protected areas forms the single most important component of a strategy for wild dog conservation. As human populations rise, pressure on wild dogs will increase. Under these circumstances, protected areas will become some of the few areas where threats to wild dogs can be minimized in the long term.

As discussed in Chapters 4 and 5, in most cases only the very largest reserves will provide adequate protection for wild dogs. There are two reasons for this. First, since wild dogs live at extremely low densities, only very large areas can sustain populations large enough to be potentially viable. Second, wild dogs frequently range outside reserve boundaries, where they encounter high-speed traffic, snares, persecution and domestic dog diseases. This means that they experience substantial edge effects, even in reserves which are large by other standards (1,000km²-5,000km²). Only very large reserves (>10,000km²) can provide core areas where wild dogs will be protected from hazards on the borders. For this reason, any measures which lead to the expansion and stabilization of protected areas - such as establishing cross-border parks, linking reserves with corridors, maintaining buffer areas around national parks, and encouraging land use favourable to wildlife on reserve borders - will make substantial contributions to the conservation of wild dogs. Such measures have been proposed or implemented in a number of areas. For example, the Niokolo-Koba National Park in Sénégal has recently been linked with Badiar National Park in Guinea, and plans have been put forward to link Kruger National Park, South Africa, with Gona re Zhou National Park, Zimbabwe by establishing further protected areas in neighbouring Moçambique. Programmes of this kind will benefit many wildlife species, but are especially valuable for the conservation of wild dogs. Wild dogs may, therefore, act as 'flagships' for the expansion of protected areas.

Wild dogs travel widely, with home ranges in excess of 1,000km² per pack, and daily movements of around 15km. Wild dogs living in small reserves are therefore vulnerable because, no matter where they go, they will cross the edge of a reserve and be exposed to human activity outside. In principal, fencing could protect wild dogs from threats on reserve borders, but fencing is extremely expensive. Some reserves are fenced in parts of southern Africa, but most of these are too small to sustain more than one or two wild dog packs. Nevertheless, a network of such reserves might support a metapopulation of wild dogs if they were protected from threats such as disease, and if some animals were translocated between sites periodically to maintain genetic diversity. Such intensive management is no substitute for protecting truly free-ranging wild dog populations and would, in any case, be prohibitively expensive in most of Africa. Nevertheless, such efforts will aid the conservation of wild dogs in highly fragmented landscapes where funds are available.

Other Wildlife Areas

Protected areas maintained by national or local governments are not the only places where wild dogs persist. Low human population densities and abundant wild ungulate prey also occur on private ranches, game farms and communal lands in many parts of Africa. Indeed, in Namibia, as well as parts of Botswana, Kenya and Ethiopia, there may be more dogs outside protected areas than there are inside them (Chapter 3). In other areas, such as Zimbabwe, even those dogs which 'live' in protected areas spend much of their time in the buffer zones outside of parks and reserves.

Wild dogs were extirpated from most private ranches and game farming areas earlier this century. However, many farmers that persecuted wild dogs to protect their stock also eradicated lions and hyaenas. Thus private land has the potential to provide ideal habitat - combining abundant prey with very low densities of competitors - if persecution could be curbed. Similar reasoning has led to suggestions that private land might play an important rôle in the conservation of cheetahs (Laurenson 1995).

An attempt is underway in South Africa to use private land for wild dog conservation: staff from the Natal Parks Board are negotiating with farmers to allow the wild dog population in the Hluhluwe/Umfolozi Park to use game farms surrounding the park (A. Maddock pers. comm.). The success of this programme will depend upon the goodwill of the farmers, and should greatly increase the possibilities for long-term persistence of this small population. Elsewhere in South Africa, however, farmers are less accommodating: when a pack of wild dogs appeared on private land along the Limpopo, local farmers immediately attempted to shoot them (M.G.L. Mills pers. comm.).

It is also possible to protect wild dogs on communal lands. For example, an innovative new programme of carnivore conservation, with extensive involvement of local people, has recently been set up in north-eastern Namibia (P. Stander pers. comm.). Outside of protected areas, persecution and disease will represent the greatest threats to wild dogs. Effective wild dog conservation will therefore depend upon minimizing these threats. We discuss the measures necessary in the next sections.

Controlling Human-induced Mortality

Persecution

Persecution is a major threat to wild dogs, especially those living outside protected areas. Most persecution is carried out by livestock and game farmers who consider wild dogs a serious threat to their stock. As discussed in Chapter 4, wild dogs may be blamed for more livestock losses than they actually cause. Where this is the case, local education will help to limit persecution - but it must be recognized that wild dogs do occasionally cause substantial losses, especially in areas where small stock (sheep and goats) are kept. Experience with wild dogs and other predators indicates that several measures can help to mitigate the problem.

Legal Protection and Zoning

Although wild dogs are classified as 'endangered' according to the IUCN threat criteria (Baillie & Groombridge 1996), the degree of protection conferred by local legislation varies among different range states. In several countries, wild dogs are only partially protected (Table 6.1); this means that, under certain circumstances, legal persecution of wild dogs can continue. For example, the government of Cameroun licensed professional hunters to shoot 65 wild dogs in the season December 1994-May 1995 (H. Planton pers. comm.). We are not aware of the numbers of wild dogs actually shot by hunters in that season - but it is extremely unlikely that Cameroun's small wild dog population could sustain the degree of persecution permitted by law. In circumstances of this kind, better legal protection represents a crucial first step towards effective wild dog conservation. We must emphasize, however, that legal protection represents only a small part of wild dog conservation: total protection failed to prevent the extinction of wild dogs in the Republic of Congo, Nigeria and Rwanda (Table 6.1).

Despite the need for better legal protection in some areas, efforts to limit persecution must take a realistic view of the threat to farmers' livelihoods. Even in livestock areas, wild dogs usually feed on wild ungulates, but they can occasionally cause substantial livestock losses (Chapter 4). Local governments may decide that large predators simply cannot be tolerated in some areas used for raising livestock, and designate such regions as predator control zones. Such 'zoning' has been an important component of wolf recovery plans in North America (Fritts et al. 1992; Mech 1995). As an example, wild dogs are sighted occasionally in agricultural areas of east-central Zimbabwe and northern South Africa, where wild prey have been depleted. It is unlikely that viable wild dog populations could persist in such areas - intensive legal protection of wild dogs might therefore alienate farmers from local conservation authorities, and could even interfere with the smooth running of other local conservation programmes (Stander 1991).

Designation of predator control zones must, however, take into account the conservation value of 'vagrant' wild dogs. As discussed in Chapter 5, movement of animals between populations - even if it occurs only occasionally - can dramatically reduce the probability that small populations will become extinct. Vagrant animals may, therefore, contribute to the long-term persistence of local wild dog populations. For this reason, farmers should not be given carte blanche to persecute wild dogs, even inside areas designated as predator control zones. Wild dogs entering such areas should be removed only if they cause serious livestock losses, and then only by local conservation authorities. Strategies for dealing with such 'problem animals' are discussed below.

Livestock Husbandry

While governments may designate some areas as predator control zones, elsewhere local conservation policy will aim to allow wild dogs to persist in areas also inhabited by livestock. Such circumstances are likely to occur on the borders of reserves inhabited by wild dogs, and in communal and private lands supporting a mixture of wildlife and livestock. In these areas, a number of measures will help to reconcile the requirements of wild dog conservation with the needs of local livestock farmers.

Better livestock husbandry may help to protect livestock from wild dogs, as well as from other predators. Masai herdsmen interviewed at 20 manyattas in a group ranch near the Masai Mara, Kenya, had no recollection of losing sheep, goats, cattle or donkeys to wild dogs, although a large pack was using the area at the time (Fuller & Kat 1990). In this area, livestock were tended continually by people and guard dogs during the day, and kept in bomas at night. Similar husbandry techniques are used traditionally to protect livestock from wolves in Italy - studies have shown that wolves often approach the bomas but rarely attack (Boitani 1992).

Wild dog predation is likely to be a more serious problem when stock are kept in large herds and poorly tended - this is certainly true of wolf predation (Boitani 1992). However, little is known about the circumstances when wild dogs kill livestock, and more research is needed before better husbandry techniques can be devised. In particular, the value of guard dogs in protecting livestock must be traded off against their role as reservoir hosts for diseases which threaten wild dogs (see below). Any programme which encouraged the use of domestic dogs as guards would also have to involve provision for disease control in the domestic dog population. Such disease control could also benefit local people since rabies, the most serious threat to wild dogs, is also a threat to people and their livestock.

Altering the type of livestock kept may also ease coexistence of wild dogs and people. Cattle are less vulnerable than sheep and goats, and there is speculation that 'traditional' cattle breeds might be better equipped than more modern breeds to deal with attacks from predators - when threatened they tend to show defensive behaviour like that of wild cattle species, protecting vulnerable calves inside of a ring of adults (G. Rasmussen pers. comm.).

Compensation Schemes

Compensation schemes have helped to resolve some conflicts between livestock farmers and wolves in North America and in Italy (Fritts et al. 1992; Mech 1995). Such schemes could be useful in wild dog conservation, especially on the borders of reserves holding important wild dog populations. The viability of compensation schemes would, however, depend upon:

1. The availability of funds to provide compensation for livestock lost to wild dogs. On some reserve borders, profits derived from tourism within the reserve could be used to fund such compensation schemes. However, many wild dog populations occur in remote reserves which generate rather little tourist revenue. In such circumstances other funds would be needed to finance compensation schemes.
2. Establishment of local standards of good husbandry to ensure that compensation does not become a substitute for adequate care for livestock (Fritts et al. 1992).
3. The availability of skilled staff to investigate alleged attacks as soon as they occur, to determine whether wild dogs were indeed responsible, and whether local standards of good husbandry had been practised (Fritts et al. 1992).
4. Adequate supervision of staff carrying out the investigation.

If all of these conditions were met, compensation schemes might form a useful component of wild dog management in some areas where they take livestock occasionally. Such schemes would, however, be expensive and should only be implemented as part of integrated management programmes including local education, work on husbandry practices and, perhaps, disease control in domestic dogs. Programmes of this kind need not, however, be aimed purely at wild dogs - several carnivore species could certainly be managed simultaneously as part of the same scheme.

Control of Poisons

Local education and compensation should help to mitigate persecution where this is directed at wild dogs specifically - for example where livestock farmers shoot them. In some areas, however, persecution is applied indiscriminately to predators in general, by laying out poison baits or adding poison to water holes. Better legal control of poisons in such countries would help to protect wild dogs - this was an important component of successful wolf conservation measures in Italy (Boitani 1992).

Problem Animals

Although wild dogs usually ignore livestock, they can occasionally cause severe problems: for example, a group of wild dogs in Laikipia, Kenya killed 66 merino ewes and 67 lambs in a nineteen-week period in 1996 (M. Dyer pers. comm.), forcing the Kenya Wildlife Service to capture the animals responsible (R. Kock pers. comm.). The circumstances under which wild dogs start to take large numbers of livestock are not clear. However, such attacks may lead to substantial economic losses which farmers cannot be expected to tolerate.

If (as in Laikipia) investigation by local conservation authorities shows that wild dogs are indeed causing losses which are both serious and sustained, and if no compensation scheme is in place or if the losses are too great to be sustained by a compensation scheme, the only solution may be to remove the 'problem animals' from the area. The first possibility is to translocate the problem dogs elsewhere. Translocation would be the best solution if suitable release sites were available. Such sites would have to:

1. Have suitable habitat for wild dogs, but no resident wild dog population
2. Have adequate protection for the translocated wild dogs
3. Be largely free of livestock. Some lions are known to develop a 'taste' for killing livestock (Stander 1990). If the same pattern occurs in wild dogs, then translocated animals might continue to present problems if they were moved to areas where they still came into contact with livestock.

Translocation is discussed in detail in Chapter 7 - however, in practice suitable reintroduction sites are very uncommon, especially outside of southern Africa. Alternatively, problem animals might be taken into captivity, where they could play a very important rôle in public education, and in conservation-related research, such as the testing of vaccines. Again, this possibility is discussed in the Chapter 7.

If no suitable sites were available for translocation, and if no captive facilities had any use for additional wild dogs, then the very last resort for dealing with problem animals would be to shoot them. This is to be avoided wherever possible. However, conservation authorities must make occasional compromises: past efforts to force people to tolerate large carnivores on their land have led to bad relations between conservation authorities and local people (Stander 1991), and a willingness to shoot wolves where they cause genuine problems has been an important component of effective wolf conservation (Mech 1995). By dealing with problem animals, wildlife authorities establish credibility with local citizens, thus improving their ability to effect conservation.

Snaring

Snares are rarely set to catch wild dogs - in most cases they are caught by accident in snares set for wild ungulates. Thus, the best way of protecting wild dogs is to invest in better control of illegal snaring inside protected areas and on their borders- this is a priority for the conservation of other wildlife in virtually all of Africa's protected areas. To be effective, however, enforcement of laws which prohibit snaring should be complemented by programmes which offer people alternative ways to secure protein, such as managed game cropping, better animal husbandry or construction of fish ponds.

Some wild dogs fitted with radio-collars have 'worn' snares unharmed, since the collar prevents the snare from strangling them (J.R.G. Unpublished data). This led to the design of an anti-snare collar which helps wild dogs to remove snares without harming themselves (G. Rasmussen pers. comm.). However, the threat posed by snares rarely warrants immobilizing animals solely in order to fit them with such collars. Investing in better anti-poaching patrols to control snaring is a more appropriate strategy, since it will provide better protection for both wild dogs and their prey.

Road Traffic Accidents

Road traffic accidents are a major cause of wild dog mortality in some areas, especially where tarmac roads pass through areas of relatively high wild dog density (e.g. Hwange National Park, Zimbabwe, Kafue National Park, Zambia, and Mikumi National Park, Tanzania). New high-speed roads should not, therefore, be routed through protected areas or along their borders - this is also a priority for the protection of other wildlife. Where such roads are already in use it might be possible to negotiate with highways departments to reduce speed limits. Road signs may also be erected along these roads, asking motorists to slow down to avoid wildlife - this has already been done near Hwange National Park. One wild dog project has built reflective tape into the collars fitted to study animals to make them more visible to motorists (G.Rasmussen pers. comm.) - however, the value of such collars in protecting wild dogs from road traffic accidents has not yet been established. At this stage it does not seem reasonable to immobilize wild dogs solely in order to fit them with reflective collars.

Managing the Threat of Disease

As discussed in Chapters 4 and 5, canid diseases represent a very serious threat to wild dog populations. In the long term, the success of wild dog conservation programmes will depend in part upon their ability to control the diseases to which wild dogs are susceptible.

Wild dogs' vulnerability to disease - and thus the need for disease control - will vary depending upon the population and disease concerned. For example, rabies causes very high mortality and represents a serious threat to all but the largest wild dog populations (Chapters 4 & 5). In contrast, since parvovirus is believed to threaten only small populations (Chapter 5), control of this disease might be inappropriate in larger populations. The threat posed by canine distemper is more difficult to assess - wild dogs have died from canine distemper in Botswana, but survived contact with the virus elsewhere (Chapter 4). Anthrax has little effect upon wild dogs in most areas (Chapter 4). It would be unrealistic, therefore, to invest large amounts of money in protecting wild dogs from anthrax unless an epidemic was believed to be threatening a particularly important population.

Since wild dogs live at such low densities, diseases which cause substantial mortality are unlikely to persist in their populations (Mills 1993). Instead, wild dogs are believed to contract diseases from reservoir hosts living at higher densities. There is good evidence to suggest that domestic dogs provide this reservoir for canid diseases in several areas. Elsewhere, wildlife species such as jackals and bat-eared foxes may act as reservoirs (see Chapter 4).

This information points to several strategies that could be adopted to protect wild dogs from disease. Attempts could be made:

1. to minimize contact between wild dogs and reservoir hosts.
2. to eradicate disease from reservoir host populations.
3. To vaccinate wild dogs directly.

Each strategy has advantages and disadvantages, depending upon both the disease concerned, and the local circumstances. We shall discuss them in order.

Minimizing Contact between Wild Dogs and Disease Reservoirs

It is rarely possible to prevent all contact between wild dogs and reservoir hosts carrying diseases that threaten them. Indeed, for anthrax, which is carried by wild dogs' ungulate prey, this would be entirely impossible. However, it would be possible to reduce contact between wild dogs and domestic dogs, mitigating the threat of disease transmission. Below, we discuss several measures which would be needed. As well as minimizing contact between wild and domestic dogs, all of these measures would also help to increase the efficacy of concurrent vaccination programmes for domestic dogs.

1. Neither tourists nor park staff should be permitted to bring domestic dogs into protected areas where wild dogs occur. If such a total ban were impossible, then owners should, at the very least, be required to prove that their dog has up-to-date vaccinations against rabies, canine distemper and parvovirus. Such dogs should, ideally, have been neutered.
2. Domestic dogs' numbers and movements could be controlled. Where wild dogs use areas also inhabited by people, domestic dogs may play an important social rôle - guard dogs might even be important in reducing livestock losses to wild dogs. Under such circumstances it may be unacceptable - or even undesirable - to completely remove domestic dogs from wild dog areas. Nevertheless, several measures that are often used in public health campaigns to control rabies could be implemented to reduce contact between wild dogs and domestic dogs. Domestic dogs should be tied up whenever possible this would not interfere with their activities as guard dogs if their principal rôle is to raise the alarm by barking. Owners of domestic dogs should be required to put collars on them, and all dogs without collars (and thus, presumably, without owners) should then be destroyed. Unaccompanied dogs should be shot on sight.
3. Wild dogs can be protected from domestic dogs by secure fencing. This may be appropriate for small reserves, but would be prohibitively expensive across most of Africa, and for larger reserves.

Eradicating Diseases from their Reservoir Hosts

If diseases that threaten wild dogs could be eradicated in the reservoir hosts that maintain them, then wild dogs would also be protected. Where the same diseases also threaten people (e.g. rabies, Cleaveland & Dye 1995), or wildlife species other than wild dogs (e.g. canine distemper, Roelke-Parker et al. 1996), protection could form part of larger-scale public health or wildlife disease control programmes. While the principal of eliminating diseases from their reservoir hosts may be a good one, a number of practical problems arise:

1. The reservoir host is not always known. For example, while domestic dogs appear to be the reservoir host for canine distemper in most areas, no reservoir has so far been identified in Selous (Chapter4). Efforts to control disease in reservoir hosts are doomed to failure if the wrong host is targeted. More research is urgently needed on the persistence of disease in wild carnivores, and the effect of between-species transmission on their epidemiology.
2. We currently have little information about the efficacy of attempts to protect wildlife by controlling disease in reservoir hosts - even where those hosts are domestic dogs. Both mathematical models and empirical studies have established the proportion of urban domestic dog populations that must be vaccinated in order to eradicate rabies (Coleman & Dye 1996). However, if domestic dogs coexist with wildlife species such as jackals and foxes, which live at high densities, then the wildlife may infect the domestic dogs, as well as vice versa. Whether the same level of vaccination cover will still protect the dogs - let alone both dogs and wildlife - is still unknown.
3. The epidemiology of rabies is relatively well understood, but few quantitative data are available on diseases such as canine distemper. This makes it very difficult to devise strategies for control of such diseases.

Despite these caveats, disease control in reservoir hosts could be a very effective way of protecting wild dogs from disease in the long term. More research is needed in this area to devise effective strategies for disease control. Such strategies would be likely, however, to combine controlling host population size and, ideally, mobility, with programmes of vaccination.

Controlling the Numbers of Reservoir Hosts

Perhaps the best way of managing disease in reservoir hosts will be to control their numbers. This would have two effects. First, it would reduce the rate of contact between wild dogs and reservoir hosts, lowering the probability that disease would enter the wild dog population. Second, it might reduce host population density below the threshold needed to maintain endemic disease. This point is well-illustrated by data collected on rabies in domestic dogs living on the borders of the Serengeti National Park: the infection persisted in one district where domestic dog density exceeded 5 dogs/km², but not in two districts where there were <1 dog/km² (Cleaveland & Dye 1995). Thus, reducing domestic dog density could, in principal, eradicate endemic rabies.

The feasibility of controlling the numbers of reservoir hosts depends upon the species involved: culling of wildlife reservoirs would almost certainly be unacceptable inside protected areas. However, where domestic dogs act as reservoirs, it might well be possible to control their population density. In protected areas that are inhabited only by park staff and tourists, there is no excuse for keeping domestic dogs. However, in other areas domestic dogs may play important rôles as guards and hunters. The possibilities for domestic dog control under these conditions would depend upon the opinions of local people, but a reduction in dog density - either by culling or contraception - might well be acceptable if approached with sensitivity. Such a reduction, especially when combined with vaccination and better control of dogs' movements, would greatly reduce the probability of disease transmission between domestic dogs and wildlife. Additional benefits of such a strategy include improved health of the remaining domestic dogs and reduced public health risks associated with rabies.

Vaccinating Reservoir Hosts

Contact between susceptible wild dogs and infectious reservoir hosts can also be reduced by vaccinating the reservoirs. Vaccination could be combined with control of host population size and mobility, but could also represent an alternative measure where local people value their domestic dogs very highly, or where the reservoir host is a wildlife species.

It is not necessary to vaccinate all the members of a population in order to eradicate a disease. Vaccination reduces the proportion of hosts in the population that are susceptible to infection. If this proportion falls below a certain critical threshold, hosts die from the disease, or cease to be infectious, before they can transmit the disease to new hosts, and the pathogen is driven to local extinction (Anderson & May 1985). For urban domestic dogs, both empirical studies and epidemiological modelling have established that rabies can be eradicated by vaccinating 70% of the population (Coleman & Dye 1996). The epidemiology of canine distemper is not so well understood, but preliminary modelling suggests that the critical vaccination cover might be as low as 50% (S. Cleaveland pers. comm.). This contrasts with related morbilliviruses such as rinderpest and measles, for which the critical vaccination threshold is much higher (M.Woodford, pers. comm., Dobson & Hudson 1995). Since most domestic dogs are concentrated around human settlements, these levels of vaccination cover can be attained realistically, if at a substantial cost (S.Cleaveland, pers. comm., R.Kock, pers. comm., Laurenson 1996).

Despite these predictions, the effect of a secondary wildlife host upon the epidemiology of rabies and distemper is unknown. It is possible, therefore, that a higher proportion of domestic dogs must be vaccinated to achieve eradication from the whole system. In the meantime, pilot vaccination programmes aimed at controlling rabies and canine distemper in the Masai Mara have managed to vaccinate 80% of domestic dogs (R.Kock pers. comm.). Empirical studies are urgently needed to determine whether such programmes can eradicate disease from wildlife populations.

Vaccination programmes planned for domestic dogs in the Serengeti ecosystem aim to create a disease-free belt on the borders of the protected area (S. Cleaveland pers. comm.; R. Kock, pers. comm.). The width of the belt in which domestic dogs must be vaccinated to protect wild dogs depends upon the mobility of both species. Since wild dogs are known to range over large areas, they could pass through the belt and encounter canid diseases outside. Trial vaccinations around the Masai Mara have produced a belt 15km wide, but this is much less than the distance that wild dogs may cover in the course of a single day. Furthermore, little is known about the mobility of domestic dogs - if migration in and out of the vaccination zone is commonplace, then the area will not remain free of disease for long. More research is needed to determine the width of cordon sanitaire needed to protect reserves from invasion by canid diseases.

Where diseases that threaten wild dogs are maintained in wildlife reservoirs, vaccination is more problematic. Oral vaccination programmes have been used routinely to control rabies in wild carnivores in Europe and North America (Wandeler 1993). Research is underway to devise similar strategies to control rabies in jackals in Zimbabwe, but has not yet reached the stage where oral vaccination could be carried out in protected areas: though effective for jackals, the virus strains used have proven highly pathogenic to some other wildlife species (Bingham et al. 1995). Other vaccine strains are available but have not yet been tested - thus, at present it would not be possible to protect wild dogs from rabies by oral vaccination of other wildlife species. Nevertheless, it is highly likely that this will be possible in the future. No such programme could be devised for canine distemper at present: the wildlife species in which the infection persists are not known, and live vaccines against canine distemper are pathogenic to several wild carnivore species (including wild dogs themselves).

Finally, it is possible that controlling the diseases to which reservoir hosts are susceptible might lead to an increase in their numbers. Endemic canine distemper caused 3-5% of domestic dog mortality in Copenhagen in the 1950s (Gorham 1966). If removing this mortality led to population growth, each annual vaccination round would become more difficult and more expensive. Furthermore, if vaccination were halted - perhaps due to lack of funds - the population of susceptible reservoir hosts would be larger, making any subsequent epidemic more severe and increasing the threat posed to wild dogs. Ongoing research on domestic dogs in the Serengeti and the Masai Mara, as well as in Ethiopia, will help to determine whether vaccination programmes do lead to such an increase in domestic dog numbers. Wherever possible, vaccination of domestic dogs is best combined with control of their numbers.

Vaccinating Wild Dogs Themselves

The most direct way of protecting wild dogs from disease is to vaccinate them. Such vaccination does, however, entail a number of problems:

The Availability of Suitable Vaccines

The safety and efficacy of vaccines against the diseases that threaten wild dogs are often unsatisfactory. Inactivated rabies vaccines have caused seroconversion in some free-ranging and captive wild dogs (Gascoyne et al. 1993), but others have failed to seroconvert (Visee 1996), or failed to establish sustained immunity (G.R.Thomson, pers. comm.; P.W. Kat, pers. comm.). At least some free-ranging wild dogs which have been vaccinated against rabies have subsequently died of rabies (Kat et al. 1995; Scheepers & Venzke 1995). The failure of rabies vaccinations to prevent rabies deaths in wild dogs has led to substantial controversy in both the scientific and popular press (see, for example, Burrows 1992; Dye 1996; Heinsohn 1992; Macdonald et al. 1992; Morell 1995) - it has been suggested that, far from protecting wild dogs, vaccination might have hastened wild dogs' deaths. This issue is discussed in detail in Appendix 1; in summary, while inactivated rabies vaccines are unlikely to have caused the deaths of the wild dogs from rabies, they also failed to prevent those deaths. The most likely explanation is that the single dose of vaccine given to each dog was not sufficient to trigger a fully protective immune response: two or more doses have been shown to provoke a better response in both wild dogs (G.R.Thomson, pers. comm.), and domestic dogs (Sage et al. 1993). More research, on captive animals, is needed to assess the efficacy of various rabies vaccination protocols for wild dogs (Chapter 8).

Problems also arise with vaccines against canine distemper. While modified live vaccines have brought about seroconversion in some cases (Spencer & Burroughs 1992), in others they have either failed to produce protective antibody levels (van Heerden et al. 1980) or have induced distemper and death (Durchfeld et al. 1990; McCormick 1983; van Heerden et al. 1989). Vaccine-induced distemper can be avoided by using killed vaccines, but studies on captive maned wolves, bush dogs, fennec, kit and crab-eating foxes indicate that such vaccines rarely cause seroconversion (Montali et al. 1983). Thus, at present there are no vaccines against canine distemper suitable for use in free-ranging wild dogs.

Modified live vaccines against parvovirus have brought about seroconversion in captive wild dogs (Spencer & Burroughs 1990).

Locating Wild Dog Packs

In order to vaccinate wild dogs in the field, one must first find them - and this is extremely difficult without the aid of radio-collars. In Selous, where wild dogs occur at high density, researchers spent the first five months of the project just looking for wild dogs (Creel 1996). Thus, vaccination would be extremely labour-intensive in areas where wild dogs had not been radio-collared, especially in thick bush. Furthermore, vaccination would have to be repeated annually to maintain immunity in adults, and to protect each new litter of pups. For this reason vaccination of wild dogs would not just be a question of paying for vaccines: vehicles, petrol and skilled manpower would also be necessary.

Halting Selection for Disease Resistance

Since a vaccination programme prevents most animals from being exposed to disease, it will weaken natural selection for disease resistance. Thus if vaccination were to be discontinued, the population would, on average, be more susceptible to infection than it had been before the programme was started. For this reason, once a vaccination programme is commenced, it may be necessary to continue it indefinitely (Hall & Harwood 1990). While there is little evidence of natural resistance to rabies, a fairly high proportion of wild dogs may survive exposure to canine distemper virus perhaps indicating some natural resistance to the disease (Chapter 4). More research is needed on the pathogenicity of canine distemper virus in wild dogs.

Choosing the Best Strategy for Disease Control

None of the options that we have discussed provides a completely satisfactory solution to the problem of disease control in wild dogs. In every case, our knowledge is limited and further research is urgently needed. Nevertheless, it is possible to suggest some circumstances in which each management strategy - or no action at all - would be most appropriate. The questions that must be answered before designing local strategies for disease control are summarized in Figure 6.1.

1. If a particular disease threatens people, livestock or wildlife species in addition to wild dogs, then controlling the disease in its reservoir host will be more appropriate politically, socially, and economically than vaccinating wild dogs directly.
2. If a wild dog population was known to be be facing an acute disease risk - for example, if the wave front of a rabies epidemic was approaching - then vaccinating wild dogs themselves might represent the most appropriate action providing the epidemic had not yet reached the wild dogs. Far from providing protection, rabies vaccination of animals immediately before they contact rabies virus may hasten the course of the disease (the 'early death' phenomenon, Clark et al. 1981). Thus the appropriate response to an acute disease risk requires very accurate information about the threats involved.
3. Smaller populations - including those re-established by reintroduction - will be more vulnerable to disease and will therefore require more active management. Disease control programmes will also be more effective if implemented over smaller areas.
4. Larger populations should be sufficiently resilient to recover from periodic disease outbreaks and may require no active management beyond monitoring of disease threats.

Conclusions

The considerations discussed above indicate that different wild dog populations face different threats, and that the appropriate management strategies will vary accordingly. In summary, though, we envisage three paradigms for wild dog management:

1) Large populations

The maintenance of large populations in extensive (>10,000km²) protected area networks remains the highest priority for Africa-wide wild dog conservation. The value of populations such as those in Selous and Kruger cannot be stated too highly. Such populations are likely to be large enough to persist in the face of even fairly dramatic perturbations, and should not require intensive management. This is fortunate, since intensive management over such large areas would be logistically difficult and extremely expensive. Protecting such populations is essentially a question of protecting their habitat: maintaining reserve integrity, controlling poaching of prey species, and avoiding the building of high-speed roads - all of these are routine components of reserve management in Africa. Given such protection, there is no reason why wild dog populations should not persist for many centuries in large reserves.

2) Smaller protected areas, reserve borders and landscape management

Where wild dogs use areas inhabited by people on reserve borders, in buffer zones connecting smaller reserves, or in areas which are not close to a protected area, their populations are likely to be under threat from persecution and disease. Managing these landscapes for wild dogs (or for wildlife in general) may be difficult, but in many areas such management is critical: in Kenya, for example, there is not a single population of wild dogs thought to be restricted to protected areas, and, while much of Kenya's wildlife exists outside protected areas, it is uncertain how long this pattern will persist with increasing population and economic growth.

In regions where management of reserves is combined with wildlife management outside protected areas, integrated carnivore management programmes should be implemented to resolve conflicts between human activity and the conservation of predators. Local conservation bodies should work with farmers to minimize livestock losses to wild dogs and other predators and might provide compensation for stock that are killed. Domestic dog populations could be controlled and vaccinated against diseases which threaten wild dogs, perhaps in collaboration with local public health authorities.

Management of this kind will be costly. On reserve borders it might be funded by tourist revenues derived from the reserve. Alternatively, external funding might be available. Such schemes will be extremely valuable on the borders of reserves holding important wild dog populations - as discussed in Chapter 4, human activities on reserve borders can represent a serious threat to such populations. The value of management schemes of this kind in other unprotected areas will depend upon local conservation policy. Intensive management is unlikely to provide value for money in areas which are intensively farmed, where ungulate prey are depleted and domestic dogs are common. It may be more useful to designate such areas as predator control zones.

3) Very small, intensively managed populations

Plans are being considered in South Africa to maintain a metapopulation of wild dogs held in a network of small fenced reserves, each containing just one or two packs. Such a metapopulation would require intensive management: for example, individuals would have to be translocated between reserves to maintain genetic diversity, and annual vaccination of wild dogs against canid diseases might well be necessary. Management of this kind will be useful, especially for the maintenance of substantial numbers of wild dogs within South Africa itself, and represents the kind of strategy that would be needed in highly fragmented habitats. However, in terms of Africa-wide wild dog conservation such schemes have a much lower priority than the continued protection or expansion of large national parks and reserves, where viable populations can persist without the need for such expensive intensive management.