BIOLOGICAL CONSERVATION Biological Conservation 88 (1999) 133±139 Livestock predation by the painted hunting dog Lycaon pictus in a cattle ranching region of Zimbabwe: a case study G.S.A. Rasmussen * Painted Dog Research Project, Natural History Museum, PO Box 240, Bulawayo, Zimbabwe Received 12 February 1997; received in revised form 8 December 1997; accepted 9 December 1997 Abstract The painted hunting dog Lycaon pictus, labelled an inveterate stock killer, was formerly slaughtered in large numbers and eradicated throughout most of its former range. Reports of a pack of painted hunting dogs estimated at 35 individuals, and cattle losses of 59 animals over a 2-year period in the Nyamandlovu region, Zimbabwe, were checked by ®eldwork. A resident pack numbering 16 individuals was con®rmed, with a home range of 5660 km2. Authentic cattle losses due to dogs from a herd averaging 3132 amounted to 426 animals out of a total of 268 losses with a peak during the calving season but few losses during the denning period. The pack fed predominantly on impala, duiker and kudu. There were indications that some of the losses attributed to the dogs were in fact due to cattle rustling and poaching. # 1999 Elsevier Science Ltd. All rights reserved. Keywords: Lycaon pictus; Cattle predation; Cattle rustling and poaching; Nyamandlovu, Zimbabwe 1. Introduction The painted hunting dog Lycaon pictus (also known as the Cape hunting dog or African wild dog) was once widespread throughout sub-saharan Africa with populations extending into North Africa (Fanshawe et al., 1991). Their range excluded true rain forest, although they were sighted in habitat as extreme as the higher altitudes of Mount Kilimanjaro (Thesiger, 1970). However, the dogs were mercilessly persecuted throughout Africa and were eradicated from National Parks. In Rhodesia, a 10 shilling bounty was paid on surrender of the tail and the skin of the head and neck (Rhodesian Agricultural Journal, 1906). They were classi®ed as vermin from 1906 until 1975 and slaughtered as putative cattle killers. For example, during the 5 year period 1956±1961, at least 2674 dogs were destroyed (Childes, 1988), the impetus for such slaughter being fuelled by irrational prejudice (for example see Mannix, 1956). Until now, ®gures for predation on domestic livestock have been biased or unobtainable as a result of poor records (Childes, 1988) or the absence of resident painted hunting dogs in farmland. Currently, in Zimbabwe, painted hunting dogs have been expanding into the * Fax: +263-9-64019. home ranges they occupied (Rasmussen, 1997) before they were exterminated (Childes, 1988). The objective of the present study was to validate the number and movements of painted hunting dogs in a cattle ranching region, to assess accurate ®gures for stock losses when dogs are resident and to evaluate the e€ectiveness of measures to decrease depredation by this endangered predator. 2. Study area The study was conducted between April 1994 and April 1996 in the Nyamandlovu cattle ranching area of Zimbabwe (28 180 E, 19 300 S) (Fig. 1). The vegetation types (Wild and Barbosa, 1967), nomenclature (Gibbs Russel et al., 1985, 1990) and relative amounts characterising the pack's home range area (HRA) are: . Type 47, Deciduous Tree Savanna (45%), predominantly Terminalia sericea with a proportion of Burkea africana and an admixture of species to include Combretum spp. and Kirkia acuminata. Grass species were represented by Bothriochloa insculpta and Hyparrhenia hirta with Eragrostis spp., including E. cylindriformis. 0006-3207/99/$Ðsee front matter # 1999 Elsevier Science Ltd. All rights reserved. PII: S0006 -3 207(98)00006 -8 134 G.S.A. Rasmussen/Biological Conservation 88 (1999) 133±139 3. Methods Fig. 1. Map showing the relative distributions of the painted hunting dog in Zimbabwe, including the Nyamandlovu pack. . Type 19, Miombo Woodland on Kalahari Sand (40%), dominated by Brachystegia spiciformis integrading with Colophospermum mopane. Grasses included Hyparrhenia hirta, Andropogon schirensis, Aristida stipitata and Eragrostis spp. . Type 35, Dry Early Deciduous Savanna Woodland (15%), with Colophospermum mopane, with understorey of Grewia spp. Grass cover was dominated by Dactylotenium aegyptum, Aristida spp., Pogonartha squarrosa and Sporobolis pyrimidalis. Elevation varies from 1110 to 1240 m above sea level with annual rainfall averaging 500±610 mm, falling mainly in the wet season from October to March. Topography is ¯at and undulating with the Umguza River running through the southern end of the pack's home range. The Insuza Vlei, an important area for the dogs, runs for 20 km through Winterblock ranch. Suitable prey species present included kudu (Tragelaphus strepsiceros), duiker (Sylvicapra grimmea) and impala (Aepyceros melampus) (Childes, 1988; Rasmussen, 1997) on the Umpuchene, Shiloh, Umguza and Winterblock ranches, as well as eland (Taurotragus oryx), bush pig (Potamochoerus aethiopicus) and warthog (Phacochoerus aethiopicus) which are less frequently taken by dogs in Zimbabwe (Childes, 1988). The whole area was utilised for cattle, stocking rates (including calves) for the ranches concerned averaging 5.5/km2 in winter to 13.2/km2 in summer. Winterblock o€ered hunting for 6 months of the year in a 200 km2 block running along the Insuza Vlei. To collect data on the number of dogs, their movements and their home range, sightings were collated from reports received from the ranchers, their sta€ and the general public. To ensure reliability, interviews were taken where necessary and the dogs were also tracked on foot. The home range was estimated by measuring the polygon bounded by the sighting points on 1:50 000 maps. Stocking ®gures and raw data for stock losses were obtained from weekly reports kept by the ranches concerned, as required by law in Zimbabwe. Stock losses were categorised into classes to aid analysis as per Table 1. The management classes covered losses generally arising from human error or negligence. Two methods were used in order to validate these losses. First, detailed interviews without prior warning were held with every sta€ member of the ranches concerned to establish circumstances leading to a loss attributed to the dogs. They were asked whether the kill had been witnessed or just a carcass found, or whether an animal was simply missing following the presence of dogs. If the kill was witnessed, was it seen or heard? To assess whether the kill was commensurate with dog behaviour, the time of day, method of killing, amount of carcass and parts remaining were recorded together with any other pertinent information. If a carcass was located and the loss attributed to the dogs, the reason for inclusion in that category was recorded, e.g. spoor nearby or dogs seen in the area. The interviews were undertaken in order to verify as many losses as possible as it was believed that claims for stock losses often tend to be attributed to painted hunting dogs once their presence was suspected, when the true cause was often hyaenas (Crocuta crocuta), feral dogs (Canis familiaris), leopards (Panthera pardus) or poachers and cattle rustlers. (In the context of this paper cattle rustling is theft of stock by live removal and poaching is the killing of the animal in situ and taking the meat.) The interviews were also essential to verify that they were in fact referring to Lycaon, as sensu lato (the name `wild dog' in both the vernacular language and English) includes jackals, domestic and feral dogs (Pringle, 1980; Skead, 1980), and also because Lycaon had been absent in the area for many years. Secondly, the dogs were intermittently tracked on foot to record movements, kills and chases, as well as to verify stock losses. This coincided with the calving season, when most losses were being reported and was undertaken over a period of 6 weeks between February and March 1996 and again in May 1996. Fresh dog spoor was located either as a follow up of a verbal report or alternatively by systematically driving slowly in circuits around the paddocks until tracks were seen. Tracking stopped only when spoor was either too old or G.S.A. Rasmussen/Biological Conservation 88 (1999) 133±139 135 Table 1 Categories utilised to group stock loss classes within the home range of a pack of painted hunting dogs, Lycaon pictus Category Management Disease Poachers Leopard Hyaena Painted hunting dogs Unknown Natural Classes Truck damage/exhaustion, stress, breakouts, swallowing foreign objects Tuberculosis, blackleg, rabies, heartwater, gall sickness, edema Rustling, poaching Toxic plants, old age, birth complication lost. All positions were recorded using the Sony Pyxie Global Positioning System and plotted on a 1:50 000 map. Fig. 2. Categorisation and chronology of predatory losses verbatim over a 2-year period within a Lycaon home range in the Nyamandlovu farming region of Zimbabwe. Chronology: (a) ranchers unaware of the presence of six dogs; (b) dogs make ®rst veri®able kill of one calf; (c) dogs left area and went to Turk Mine; (d) pack returns and breeds, presence very evident no losses veri®ed; (e) ranch-noti®ed adult stock losses September/October not feasible for Lycaon; (f) veri®ed calf losses to Lycaon commencing December. on completion of the study 59 stock were reported lost over the 2-year period (Fig. 3). 4. Results 4.2. Field studies 4.1. Rancher-reported losses and dog numbers 4.2.1. Painted hunting dog numbers and home range The ®rst new record of dogs in the Nyamandlovu region was in April 1994 when six males were reported to the Painted Dog Research Project (PDRP). They were reported on a further 13 occasions that year, but by the end of 1994 the pack numbered ®ve, as one was killed by a car. In early February 1995 ®ve male dogs, believed to be the same ones, were sighted in the Lonely Mine area some 35 km to the north west, and at the end of February 1995 a group of eight was sighted in the same area. In April 1995 eight dogs were present in the Umguza area and from the distinctive markings of one individual and spoor of another it was evident that the ®ve males had returned with females. They bred in June 1995 and were sighted on the Insuza Vlei in September 1995. In December 1995 another individual was killed by a car and in April 1996 foot tracking and records of all sightings revealed that there was only one pack in the area, numbering 16 individuals. The home range was estimated at 5660 km2, encompassing at least nine ranches. From interviews with the ranchers it was evident that, historically, dogs had bred in the Insuza Vlei but they had been exterminated and there had been no resident pack for at least 50 years. During January and February 1995 nine calf losses attributable to ®ve dogs were reported (Fig. 2, Table 2) but only two were validated and cattle poaching was suspected as the dogs were not present for part of the period (Fig. 2). In April 1995 eight dogs reappeared and for the 6 subsequent months, during which period the dogs were denned on the Insuza Vlei, three weaners (age class 1±2.5 years old) were reported taken by the dogs, although the foreman of the Winterblock ranch reported that the dogs were feeding almost exclusively on the abundant impala in the region. In October 1995, Winterblock ranch reported that the dogs had eaten 18 full-grown cows in the 21-day period from 28 September to 19 October 1995, 14 of which were reputed to have been eaten at a rate of 1 a day, with the whole carcass except a few bones being taken. Since painted hunting dogs eat 4‹0.35 kg/dog/day, the adult stomach capacity being c. 9 kg (Creel and Creel, 1995), these losses were neither veri®ed nor credible for the Nyamandlovu pack. After a visit to Winterblock to discuss the feasibility of this, all reports of adult stock losses ceased (Fig. 2). There were no further reports until 7 December 1995, when one calf and one weaner were reported as taken. Between January and early February 1996, a 5week period, alleged losses amounted to eight weaners and 26 calves, including four calves being taken and eaten at one time. The reported losses were thus averaging one stock animal every other day. It was stated that there were 35 individuals in `at least two packs'. Following these reports an intensive enquiry started and 4.2.2. Stock losses and prey Intermittently, over a 7-week period between February and April 1996, the dogs were tracked on 12 occasions for distances of 3 to 18 km, and for periods of up to 3 days. During a tracking period totalling 17 days, known prey comprised two duiker, three impala and a kudu female. No stock was lost, even though on ®ve occasions the dogs entered paddocks containing 136 G.S.A. Rasmussen/Biological Conservation 88 (1999) 133±139 calves. On one of these occasions they ate a duiker and on another the pack simply passed through and chased kudu in the adjacent paddock; this behaviour indicated that the dogs were in hunting mode. On the third occasion they slept 400 m away from the stock, while on a further two occasions the dogs were located in paddocks where calves were situated, but none was taken or injured although on one occasion the stock was harassed. Furthermore, when the PDRP team was present from 8 to 31 February 1996, immediately subsequent to a reported `spate of calf killings', there was only one stock loss. Immediately the team left the area, four calves were reported as having been taken by dogs. Reappearance of the team, however, resulted instantly in a reduced claim, with the remaining losses being highly suspect. The other three calf losses, though recorded in the sample as dog kills, were highly suspect as burnt calf remains were found in the area. During this period the main areas being utilised by the dogs encompassed 45 815 ha where there were c. 2900 head of cattle, with a calf crop of c. 550. Two months subsequent to this study, one of the sta€ members was caught and convicted of slaughtering and selling a heifer whose killing had been attributed to painted hunting dogs. Some of the herdsmen interviewed were unable to recollect any details of a number of kills they had personally reported to management in the last 2 months. Exaggerated claims were commonplace and only indepth questioning enabled a clearer picture to be gleaned. Total losses thus amounted to a maximum of 15 calves and three weaners rather than 27 calves and nine weaners for the 102-day peak loss period from 7 December 1995 to 18 March 1996 (Fig. 3). The actual peak losses being sustained thus averaged one domestic stock animal every 5 days. Statistics on stock losses, obtained from weekly reports, combined with the actual ®ndings clearly put the impact of predators into perspective (Fig. 4). Furthermore, the major losses, both numerically and ®nancially, are represented by adult stock and the dogs are not responsible for these. In order to evaluate the losses due to dogs and other factors on a ®nancial basis, the stock classes were proportionally multiplied by the relative monetary replacement values, namely calf=1, weaner=4.7 and adult=11.7. This analysis shows that the losses from dog predation represent 2% of all the losses and 0.17% of the value of the herd (Fig. 4, Table 2). Certain ranchers insisted that the dogs were selectively targeting their farms so an analysis of the number of kills relative to the ranch size and known HRA of the pack was undertaken (Table 3). The observed kills were compared with the expected number for the area utilised. Although the null hypothesis was not formally rejected (w2=3.68, p=0.057), a larger sample might have reached signi®cance and the results were suggestive of selectively higher kills at Winterblock or possibly poorer stock management there. Interestingly, three ranches, Maraposa, Merryland and Mararoa, which had cattle only (as the game had been poached out), su€ered no stock losses, possibly indicating that these areas were not used by the dogs for hunting purposes but for transit only. Natural prey density was high according to indications from spoor, visual sightings and reports from the ranchers and their sta€. In addition, as the survey revealed that during the whole of a 2-year period the dogs consumed 426 stock animals, it was clear that they were feeding predominantly on natural prey and not selectively taking cattle as `an easy target', as claimed by the ranchers. 4.3. Measures to decrease depredation Two strategies were employed by the ranchers to reduce depredation. 1. Reliance on the herding instinct of the cattle to form a protective circle around the calves. This was monitored on two ranches; one resident pedigree Hereford herd was e€ective at reducing losses (Umpuchene) where there were nine reported incursions by the dogs over a 3-month period, with only two losses, which occurred on the ®rst two incursions. This owner reported that the cattle became more accustomed to the presence of the dogs and thus less intimidated. On one occasion the cattle had actually injured one of the dogs, Fig. 3. Reported vs actual maximum stock lost to Lycaon pictus in the Nyamandlovu region for the period April 1994 to April 1996. Fig. 4. Losses from a total of 268 head of stock within the home range of a pack of painted hunting dogs in the Nyamandlovu region of Zimbabwe over an 18-month period in 1994 to 1996 adjusted for circumstantial evidence. Numbers and against the columns show the percentage losses after adjustment for relative values of adults, weaners and calves. 137 G.S.A. Rasmussen/Biological Conservation 88 (1999) 133±139 Table 2 Stock losses to painted hunting dogs Lycaon pictus as a percentage value of the total herd, averaging 3132 head carried within the pack's home range during the period October 1994 to March 1996 Verbatim Adjusted Management Disease Natural Unknown Poached Strays Hunting dogs Hyaena Leopard 4.08 4.08 2.05 2.19 1.43 1.57 0.38 0.56 0.12 0.32 0.06 0.20 1.01 0.17 0.11 0.15 0.07 0.07 Total losses=9.31 Table 3 Analysis of stock losses in relationship to the home range of a pack of painted hunting dogs in the Nyamandlovu region of Zimbabwe over a period of 18 months in 1994 to 1996 Ranch Area utilised (km2) Expected kills Observed kills Winterblock Umguza Bletchingley Shiloh Umpuchene Maraposa Merryland Mararoa Mzohluzo Totals 300 143 9 73 > 52> > 25= 23 217 20> > > 14; 10 660 11.8 5.6 9 2.9 > 2.1 > > 1.0 = 0.9 8.6 0.8 > > > 0.5 ; 0.4 26 15 7 1 0 2 0 0 0 1 26 9 > > > = 4 > > > ; which shortly afterwards was found dead. On the second ranch (Winterblock), where the stock was predominantly brought in for fattening and thus was perhaps not so well acclimatised to the region, there was no anti-predatory behaviour reported and higher losses were sustained (Table 3). 2. Use of a kraal. One rancher had a visually `open' kraal (constructed of poles only), which was constantly manned and close to habitation. This lost no stock. Another rancher (on a 15 000 ha leased section of Winterblock where proportionally most of the losses occurred) had an `open' kraal far from habitation. This was infrequently manned and lost ®ve calves over 5 weeks. Subsequent visual `closing' of the boma with brushwood prevented further losses until the lessee moved. Clear line of sight is evidently important to painted hunting dogs in farmland as they were never observed to cross a fence line from one paddock into another when visibility was in any way restricted by scrub or stands of grass growing along the fence line. Such caution has not been observed with painted hunting dogs on similar exercises by the author when tracking in National Parks or in other areas where the dogs are not in any way harassed. 5. Discussion The vernacular name painted hunting dog merits some explanation. The ®rst name denoted to the species from a type specimen in Mozambique was the Cape hunting dog (Temminck, 1820). In that era the name Cape was applied irrelevantly to a number of species. Modern protocol indicates that a place name should only be utilised where endemism exists or such a name occurs in the speci®c epithet. `Hunting dog' accurately describes the ecological role of the species, particularly as they possess a unique molar/premolar con®guration that improves carnassial shear and is a diagnostic characteristic for the genus. On the carnassial, the inner cusp of the talonid is missing, providing not a basin for crushing but an additional cutting blade. The ®rst upper molar is correspondingly modi®ed, both features indicating the highly predaceous habit of the species (Ewer, 1973). Consequently all Lycaon fossil species have been called hunting dogs, e.g. L. africanis (African hunting dog), in accordance with the speci®c epithet, and L. atrox (Kromdrai hunting dog). Two undescribed intermediate fossils from Sterkfontein are labelled L. sp. indet. (hunting dog species undetermined) yet the label `wild dog' is ascribed to the modern animal. Lycaon pictus translates as `painted wolf-like animal', which correctly indicates a similarity in appearance only with true dogs (Canis spp.), but recognises the uniqueness of the genus. The names wild dog or African wild dog falsely imply taxonomic anity to Canis, so if any species in Africa is to be called the African wild dog it should be the jackal as it is at least Canis (Pringle, 1980; Skead, 1980). The vagary is further complicated in literature when one is referring to wild versus captive Lycaon. The name wild dog developed during the era of persecution of all predators when the name applied to feral dogs, hyaenas, jackals and Cape hunting dogs (Pringle, 1980). `Painted', aside from being a direct translation of the speci®c epithet, accurately describes the unique varicoloured markings of each individual. Apart from being misleading, continued use of the name wild dog does little more than further fuel negative attitudes and prejudice, which are detrimental to conservation e€orts. 138 G.S.A. Rasmussen/Biological Conservation 88 (1999) 133±139 The con¯ict between predators and farmers is here to stay, so it is essential to achieve a better understanding of the impact of predators particularly endangered predators, on livestock. From this study three things were clear. First, the presence of predators provides a convenient and plausible excuse for herdsmen to explain missing stock and hide the occurrence of cattle poaching and rustling, which claim far more stock in Zimbabwe as a whole than predators (according to the Zimbabwe Cattle Producers' Association database). Unfortunately a disproportionate amount of both energy and needless killing is exacted in the name of `predator control'. As painted hunting dogs are diurnal, highly mobile, and move in packs with enormous home ranges, they leave a great deal of spoor and are thus `highly visible'. As a result their `culpability' is often related more directly to this `visibility' than to direct evidence. In contrast, most of the other large predators (with the exception of the cheetah, Acionyx jubatus) are nocturnal, have smaller home ranges and are thus less conspicuous. Furthermore, reports showed that prior to the known presence of the dogs, leopards and hyaenas were considered responsible for the stock losses. However, with the known arrival of the painted hunting dogs all subsequent predator losses were attributed to them (Fig. 2). This can be attributed to an intrinsic loathing for `wild dogs' by farmers who put prejudice before rationality. Evidence for this statement is that, out of 59 supposed kills by the dogs, veri®cation of sta€ reports by the ranch owners/managers was only undertaken on two occasions. In addition with large packs, particularly those allowed extended carcass access time (salient in farmland situations where there are fewer major kleptoparasitising predators), dogs can leave few or no remains. This is especially true when prey is the size of calves, duiker and yearling impala. These are matters that the owners themselves must investigate to obtain a true picture of stock losses. Second, correct stock management clearly has the ability to minimise losses due to Lycaon. Resident, welladapted cattle, especially indigenous breeds, have a far greater anti-predator herding instinct, which in freeranging stock will minimise losses (Middleton, pers. comm.). Another useful deterrent is the use of a visually closed kraal. This prevents entry as farmland painted hunting dogs are particularly wary of entering an area where there is reduced visibility. In Zimbabwe, this behavioural aspect is used by poachers intent on snaring the dogs (in certain areas of Zimbabwe Lycaon body parts are utilised for traditional medicine). The poachers block o€ areas with brush wood, thus creating `visually open' channels guiding the dogs into snare lines. Conversely, a kraal neither manned nor with brush sides is courting problems as once predators are inside they are more likely to take higher numbers of stock as the herd is con®ned and thus easily captured. All the above sys- tems were employed by various ranchers in the study site with the predicted results. A number of ranchers argue that free ranging is more economical from a stang point of view and that kraaling limits grazing access time, therefore it is essential to remove predators. This is not necessarily true. For example, a 5-year trial study in northern Rhodesia comparing free ranging to herded cattle found no discernible di€erences between the two groups (G. Fredricks, pers comm.). In addition, Hilton-Barbour (1997) argues that while there are additional stang costs, there are advantages to the herding system because increased vigilance and inter-herd proximity justify utilising additional manpower, stock theft is reduced, calves show increased weaner weights because they are with their dams more, sick animals or cows with calving diculties are noticed earlier, bulling power is doubled, infertile bulls can be identi®ed earlier by noticing if the cows recycle, cattle become more manageable with the continual handling thus reducing losses to physical injuries from breakouts, as well as bene®ting from fewer stress-related problems. Finally, as in this case study, losses due to Lycaon can only be deemed negligible over the time-frame and area concerned. The argument by ranchers that the loss of a calf or weaner represents a potential loss of an adult animal and thus must be ®nancially equated as such cannot be upheld as both calf and weaner stock are replaceable. In this study, certain ranchers commented that the presence of painted hunting dogs stresses the herd in a way that other predators do not, and this needs investigation. Whether this is related to lack of familiarity or whether the dogs have a natural strategy to create stress in their prey and so facilitate herd breaking, can only be surmised. While stock losses from any cause are lamentable, a disproportionate amount of time, energy and money can be spent trying to resolve the wrong problems as a result of prejudice against predators, particularly `wild dogs'. Clearly this prejudice runs much deeper than ®nancial loss. For example, in this particular case study more value in stock was lost as a result of cattle swallowing items such as plastic bags, bailer twine and in one case a shoe (36% of the management loss class) than was lost to the dogs. In fact dealing with the more dicult and dangerous problem of poaching and rustling, as well as concentrating on reducing losses from the management and disease categories, would be ®nancially more productive than predator control. As a percentage of the total stock present, the proportional losses to dogs were minimal and did not justify the relative furor created by the presence of the dogs. What was notable and regarded as an anomaly was the almost non-existence of stock reported lost to poaching and rustling. In Zimbabwe cattle poaching and rustling is a widespread industry; the ®gure of only G.S.A. Rasmussen/Biological Conservation 88 (1999) 133±139 0.12% is unrealistically low and well below any national averages. In Zimbabwe on major commercial farms alone, c. 160 head of cattle are known to be rustled every week (according to the Zimbabwe Cattle Producers' Association database). This provides further evidence for the fact that a proportion of the above losses attributed to dogs, leopards, hyaenas and unknown predators were in fact poached or rustled (Table 2). It is signi®cant that the moment the dogs were reported as killing stock all predator losses were then attributed to painted hunting dogs (Fig. 2). However, statistics alone will not prevent the dogs being shot on the grounds of being a predator. Compensation is always called for by ranchers but the high proportion of unproven predator losses would undermine any sensible compensation scheme. Compensation schemes can also create administrative problems (Cozza et al., 1996). Such a scheme would require mobile manpower to verify each claim, and claims that were not met or veri®ed on time would cause friction between interested parties. Nevertheless there is a strong argument for a scheme in key zones adjacent to National Parks. Packs of dogs operating in the boundaries of unfenced National Parks are highly vulnerable to becoming victims of the `vacuum e€ect' (Rasmussen, 1997). Consequently a 30 km deep `cordon sanitaire' for dogs is highly desirable to increase the e€ective safe area for packs inside national parks (Rasmussen, 1997). Another possible approach in such zoned areas, could be for ranchers to insure valuable stock against predation, particularly with pedigree herds. Premiums could then be re¯ected by parameters such as management strategies and risk of predation due to proximity to the wildlife area. As a result, rather than `managing the predator' (Mills, 1991), in order to meet regulations set by insurers, such a system would encourage ranchers to adopt an active herd management strategy. Chronic negative attitudes are slow to change and therefore may not be altered in time to save the painted hunting dog from further decline. As this is a long-term problem, with few painted hunting dogs remaining in Africa and a depauperate gene pool (Girman et al., 1993), translocation of packs would seem the best shortterm solution. The technique for capturing a whole pack exists (English et al., 1993) but there is a need to research subsequent survival before it can be adopted on a larger scale. The ®rst step is to search for suitable translocation sites on a pan-African basis so that the groundwork is done before the need arises. Further case studies should also be undertaken to assist in formulating a management strategy for handling and analysing future problems. Ultimately the conservation of predators in farmlands must come `from within', with public relations and the education of farmers being essential for Lycaon as well as for other predators. 139 Acknowledgements I would like to thank the Department of National Parks and Wildlife Management Zimbabwe for their assistance and permission to research this endangered species. Thanks also to Dr. M.G. Mills for his valuable comments on the ®rst draft. References Cozza, K., Fico, R., Battistini, M., Rogers, E., 1996. The damage± conservation interface illustrated by predation on domestic livestock in central Italy. Biological Conservation 78, 329±336. Childes, S.L., 1988. The past history, present status and distribution of the Hunting dog Lycaon pictus in Zimbabwe. Biological Conservation 44, 301±316. Creel, S., Creel, N.M., 1995. Communal hunting and pack size in African wild dogs, Lycaon pictus. Journal of Animal Behaviour 50, 1325±1339. English, R.A., Stalmans, Mills, M.G.L., van Wyk, A., 1993. Helicopter-assisted boma capture of African wild dogs Lycaon pictus. Koedoe 36/1, 103±107. Ewer, R.F., 1973. The Carnivores. Cornell University Press, Itahca, NY, p. 41. Fanshawe, J.H., Frame, L.H., Ginsberg, J.R., 1991. The wild dogÐ Africa's vanishing carnivore. Oryx 25, 137±146. Gibbs Russel, G.E., Reid, C., Van Rooy, J., Smook, L., 1985. List of species of Southern African plants, 2nd ed., Part 2. Memoirs of the Botanical Survey of South Africa. Botanical Research Institute, Pretoria, p. 51. Gibbs Russel, G.E., Watson, L., Koekemoer, M., Smook, L., Barker, N.P., Anderson, H.M., Dallwitz, N.J., 1990. Grasses of Southern Africa. Memoirs of the Botanical Survey of South Africa. Botanical Research Institute, Pretoria, p. 58. Girman, D.J., Kat, P.W., Mills, M.G.L., Ginsberg, J.R., Borner, M., Wilson, V., Fanshawe, J.H., Fitzgibbon, C., Lau, L.M., Wayne, R.K., 1993. Molecular genetic and morphological analyses of the African wild dog (Lycaon pictus). Journal of Heredity 84, 450± 459. Hilton-Barbour, G., 1997. A new look at protective cattle management. The Bulletin April, 14±15. Mannix, D.P., 1956. Savage killers of Africa. The Saturday Evening Post, May 1956, pp. 45, 54, 56. Mills, M.G.L., 1991. Conservation management of large carnivores in Africa. Koedoe 34, 81±90. Pringle, J., 1980. Killers and the conservationists. T.V. Bulpin and Books of Africa (Pty Ltd). Capetown, p. 78. Rasmussen, G.S.A., 1997. Conservation Status of the Painted Hunting Dog Lycaon pictus in Zimbabwe. Ministry of Environment and Tourism, Department of National Parks and Wildlife Management, Zimbabwe June 1997, pp. 5, 6, 11, 12, 13±16, 22, 23, 32, 37, 47, 48. Rhodesian Agricultural Journal, 1906. Destruction of wild carnivora, p. 280. Skead, C.J., 1980. Historical Mammal Incidence in the Cape Province. Department of Nature and Environmental Conservation, Capetown. Temminck, C.J., 1820. Ann. Gen. Sci. Phys. II 1820±1841: Monographies de mammalogies, 2 vols. Paris and Amsterdam. Thesiger, W., 1970. Wild dog at 5894 m (19 340 ft). East African Wildlife Journal 8, pp. 202, 203. Wild, H., Barbosa, L.A.J., 1967. Vegetation Map of the ¯ora Zambesiaca Area and Supplement. Government Printer, Salisbury, Rhodesia (Harare, Zimbabwe).