The Botswana Predator Conservation Program (BPCP) is an umbrella program for all large predator conservation research projects within its domain. It has a dual mission:

• To use rigorous scientific inquiry to acquire essential knowledge regarding the behavioral ecology and communication systems of the African wild dog, cheetah, leopard, lion and spotted hyena, and to apply this knowledge to promote solutions for the preservation of Africa's large predators and their habitats.
• To link conservation and environmental issues to the decision making in the ongoing development of rural Botswana.

Aims

• To promote the sustainability of threatened and endangered carnivore species and their habitats in Botswana through research, education and management.
• To support ongoing successful management programs and to develop and apply new techniques that sustainably protect threatened carnivores in their natural habitats and ecosystems.
• To train and mentor the people of Botswana to be the conservationists and wildlife custodians of the future.

The BPCP funds and supports the logistics of all living and research requirements of a field research camp. This includes seven 4x4 research vehicles, two aircraft, operations and support staff, aerial radio tracking, and radio telemetry of the study population of wild dogs and the entire guild. A number of international graduate students, Botswana national students, and local staff make up a strong field team.

The program objective is carried out through several different core projects that fall under the umbrella of the Predator Conservation Program. These include:

Natural Science Research

• Predator Guild Study: Population Monitoring and Behavioral Ecology of African wild dogs (Lycaon pictus), leopards (Panthera pardus), cheetahs (Acinonyx jubatus), spotted hyenas (Crocuta crocuta) and lions (Panthera leo).
• Herbivore Population Monitoring
• BioBoundary Project

Graduate Research Projects

• Territoriality and Livestock Conflict
• Vocal Communication in Large Carnivores
• Spoor Surveys of Large Carnivores
• Spotted hyenas, lions, and wild dogs

Social Science Research

• Domestic Dog Management
• Coaching for Conservation

Dr. J. Weldon McNutt is the director of the program and is responsible to the government of Botswana to supervise all scientific research. Social programs are directed by Lesley Boggs. The Botswana Predator Conservation Program is academically driven and has a strong education and training component. Students from around the world have undertaken Masters and Ph.D. degrees with BPCP. BPCP initiated a Botswana national Masters student-mentoring program with the aim to match Botswana national Masters students to each Ph.D. research project. Dr McNutt is currently supervising four Ph.D. graduate research programs and one Masters study. In 2007, BPCP supported two full-time Botswana national research assistants in the program. One is now enrolled in a Masters Program at the University of Pretoria.

Click on the links below for information about our projects and research studies.

Founded in 1989, the Botswana Wild Dog Research Project is the longest running large predator research project in Botswana and one of only a handful of its caliber worldwide. Its continuous field-based research was originally designed to provide information on the behavioral ecology of African wild dogs (Lycaon pictus), Africa’s most endangered large carnivore. Research includes habitat requirements, prey species preference, prey density, dispersal patterns, reproduction patterns, and population health. The project has since evolved into an umbrella program for large predator research focusing on the effectiveness of natural resource management systems and the promotion of the management of free ranging predators.

The project was started by Dr J. Weldon “Tico” McNutt as part of his doctorate in Animal Behavior in 1989. At the time, there was virtually no information on African wild dogs from anywhere outside the Serengeti Plains. The social and community research interests of Lesley Boggs complemented this pure science inquiry when she joined the project in 1992.

“The African wild dog is Africa’s most endangered large carnivore. BPCP’s research on wild dogs has made it abundantly clear that the health and welfare of not just the wild dog population but the entire predator population is a key indication of health of the ecosystem. Our project has evolved into an ongoing program of applied research and conservation focusing on the entire large predator guild." — Tico McNutt, Director, BPCP

Today, the Botswana Wild Dog Research Project is the standard to which most other wild dog projects are based, including programs in Zambia, Zimbabwe, Namibia, Kenya, and South Africa. Most of these projects actively collaborate through the Wild Dog Working Group (a subgroup of the Canid Specialist Group) for which Dr. McNutt is the coordinator. Most of the researchers in these other projects have visited the Botswana Wild Dog Research Project and continue to collaborate through regular working group communications and meetings.

From the annual report submitted to the Government of Botswana by The Botswana Predator Conservation Program (Dr. J. W. “Tico” McNutt, Director and Project lead Gabriele Cozzi)

In this census project conducted in the eastern Moremi Game Reserve and adjacent Wildlife Management Areas for four days in October 2007, spotted hyenas (Crocuta crocuta) were attracted to a site through broadcasts of various recordings of hyena calls. The objective was to obtain a density estimate from the number of hyenas attracted to the calling stations. Calling stations were distributed in the survey area according to an estimated maximum response distance of spotted hyenas from elsewhere. This procedure has an established history for estimating hyena densities in varied habitats.

After nightfall, the team broadcasted hyena calls for six-minute periods, alternating with four minutes of silence over a 30-minute sample period. The calls used included hyena whoops, the sounds of several hyenas vocalizing at a kill, wildebeest (Connochaetes taurinus) alarm cries (when captured by a predator), vocalizations of an encounter between two hyena clans, and hyenas mobbing lions (Panthera leo). This variation was intended to maximize the number of hyenas attracted to the calling stations by including those not only motivated by a possible feeding opportunity but also those motivated by a territorial challenge.

A grid with 13 calling stations set a minimum of 7.5 km apart was laid out, covering all habitat types in the sampling area. Furthermore, BPCP used knowledge of clan territorial boundaries (established from multiple hyenas fitted with GPS radio collars) to ensure that the same clan’s territory was sampled with successive calling stations on the same night. These measures minimized the risk of double counting hyenas at different calling stations.

Using hyenas fitted with GPS radio collars, we monitored the response of specific individuals over known distances. A total of 47 hyenas were counted at the 13 calling stations, yielding a density estimate that is considerably greater than previously reported habitat specific estimates. There are a number of possible explanations for an underestimate deriving from known individuals. The most probable cause is that clan membership is considerably larger than the number of adults we have identified in the vicinity of den sites. Also contributing to an underestimate is that the analysis does not assess territory overlap among neighboring clans, resulting in a cumulative overestimate of the sampled area. Nevertheless, our results here compare well with previous studies using the same sampling method and suggest that our survey area supports a hyena density that is intermediate in the range of densities reported from other regions.

News update on the NTGR release of Wild Dogs from Craig Jackson*

On Saturday 5 April the Northern Tuli wild dogs were finally set free! Approximately 50 people attended to observe the event and the release went smoothly. The pack of 18 wild dogs including last year’s 9 pups (now yearlings) was coaxed out of the gates of the enclosure they had been living in since October of last year, using a freshly killed impala carcass, their natural staple diet. After eating the impala the dogs trotted about 300m down the road to a large and shady Mashatu tree. It looked as though they might rest there during the heat of the day, but it soon became apparent they had other ideas.
Exactly two hours and twenty minutes after gaining their freedom, the entire pack was stretched out halfway through the Limpopo River, apparently attempting to cross back into South Africa. At this point it appeared as though this was going to be an extremely short experiment.

However, the Limpopo being in full flood and flowing steadily seemed to discourage them from attempting to complete the crossing. Instead they frolicked around in the shallows for an hour and a half, and then retreated back to the Botswana side and back into the Northern Tuli Game Reserve. Then the pack gradually worked their way along the banks of the Limpopo in a southwesterly direction. Early the next morning they were located almost 10km further west, on a main road just west of Tuli Safari Lodge - but minus the three 2 year old males. We predicted prior to their release that these males would leave the rest of the pack due to their age and kinship (unrelated to the dominant male), but we had no way of knowing it would be so soon. Anyway, later that evening the dispersing three were located resting under a tree several kilometers further west than the rest of the pack - but within 100m of one of our BioBoundary scent marks. Since then they have been located infrequently. However, and notably, they have been sighted north and west near Nitani Lodge, and later, further east in Jwala, and most interestingly, appear to be very closely tracking the BioBoundary perimeter. The main pack returned over the next couple days to the place of their release, and appeared to be waiting for someone to bring them some food. On the afternoon of the 8th of April, the pack was passing through a pan called “Pete’s Pond” in Mashatu, the largest of the NTGR landholders. Here they disturbed a breeding herd of elephants that had probably not seen wild dogs in many years resulting in a serious charge by the elephants that was captured on camera (see http://www.youtube.com/watch?v=xaXazDHQ3Eg) They were no doubt getting pretty hungry by this time. Over the first few days, they were observed making feeble attempts at hunting impala with no success. However, after capturing a young warthog and then a few rabbits, they seemed to start to get back into the routine of capturing their own meals. By Friday they were observed killing an impala. They have since made several impala kills and are feeding themselves as few other large carnivores can, with great efficiency.

Map showing the Northern Tuli Game Reserve in easternmost Botswana. The border with Zimbabwe is along the Northeast. The Limpopo River, the boundary between Botswana and South Africa, is the southern boundary. As the enclosure was situated next to the Limpopo River and because the river was in flood, the southern boundary of the NTGR was not included in the layout of the scent marked BioBoundary (green dots). The area enclosed by the scent marks is approximately 360km2.

A summary of the important observations from the first week following the release of the Northern Tuli wild dog pack into the BioFenced NTGR are as follows:

1. Firstly, there is little doubt that the released wild dogs are aware of our translocated scent marks.
   • Their tendency to stop and rest several times exactly along the BioBoundary is surprising even for us.
   • The behavior of the dispersing subgroup of three males effectively traversing the BioBoundary to the north and east also strongly suggests a ranging pattern in response to the perception of a neighboring pack territory boundary.
   • The dominant pair were observed scent marking (urine) several times within a few hundred meters of the exact locations where we placed scent marks along the BioBoundary.
2. Secondly, the main pack (now 15 wild dogs) appears to be particularly interested in probing the possibility of returning to the south, presumably back to South Africa where they came from (an outcome demonstrating a remarkable ability to ‘home’ to their place of origin reported previously for wild dogs in an attempt to translocate an intact pack in Zimbabwe). If the Limpopo River were dry or lower, there is little doubt among the observers that the released dogs would have crossed immediately back into South Africa (see photograph above). Despite this predisposition however, they appear to be cognizant of and responding to their perceptions of a neighbouring pack (the BioBoundary). The ranging of the dispersing group of three 2yr old males appears much more like a new pack inspecting the extent of their new range as identified by the BioBoundary.

Stay tuned for updates …
* This report was prepared by J.W.McNutt, from reports written by Craig Jackson, the field researcher in charge of monitoring the wild dog reintroduction into the NTGR, and PhD student, University of Pretoria, South Africa. JWM (“Tico”) is the Director of the Botswana Predator Conservation Program and scientific supervisor for the BioBoundary experiment in NTGR

From the annual report submitted to the Government of Botswana by The Botswana Predator Conservation Program (Dr. J. W. “Tico” McNutt, Director)

In this project monitoring five clans of spotted hyenas known to be resident in an area of the southern part of Moremi Game Reserve and the adjacent Wildlife Management Areas, four of the five clans are monitored and identified (Table 1) through GPS collars. Four individuals in three clans (Athena, Ceres and Poseidon) are fitted with GPS radio collars, and two females in clan Ginger are fitted with a traditional VHF radio collar. Collaring of additional hyenas including at least one male and one female in the fifth clan is planned for the first semester of 2008.

Table 1.Four spotted hyena clans currently in the large predator guild research project.

In contrast to traditional VHF radio collars, GPS radio collars can be set to record an animal’s location at any time without an observer being present, providing us with the unique opportunity to remotely follow animals on a fine spatial and temporal scale. By the end of September, individuals of all five large carnivore species that occur in northern Botswana have been fitted with GPS radio collars, including five lions (Panthera leo), three leopards (Panthera pardus), two cheetahs (Acinonyx jubatus), four spotted hyenas (Crocuta crocuta), and one African wild dog (Lycaon pictus).

Location data from spotted hyenas fitted with GPS radio collars earlier this year show an interesting difference in home range size according to habitat type (Fig. 1). Clan Ceres and Clan Athena almost exclusively inhabit grassland habitat or former flood plains associated with Acacia woodland. The range of Clan Poseidon falls almost entirely within mopane woodland, which supports comparatively low prey density. Clan sizes and memberships are still being determined, but our analysis suggests that the density of spotted hyenas is likely to vary widely depending on habitat type and associated prey availability in northern Botswana.

GPS radio collars are also equipped with two bidirectional motion sensors (accelerometers) that record an animal’s activity continuously at five-minute intervals. Interestingly, activity data from spotted hyenas show that males are considerably more active than females: 10 hours versus 7.5 hours of activity per day.

Figure 1. Home ranges of three neighbouring clans of spotted hyenas in northern Botswana (Moremi Game Reserve and adjacent Wildlife Management Areas): clan Ceres in red, clan Athena in blue and clan Poseidon in green. Note the interesting loop that Ceres (orange dots) made around clan Athena’s home range within 48 hours. Recording this kind of movement patterns is only possible with GPS technology.

From the annual report submitted to the Government of Botswana by The Botswana Predator Conservation Program (Dr. J. W. “Tico” McNutt, Director)

Permission to conduct research on lions was granted to BPCP in June 2007. During the second half of the year we identified three different prides inhabiting the core study area that comprises the southern part of Moremi Game Reserve and the adjacent Wildlife Management Areas (WMA). In October 2007, five adults in three different prides were fitted with GPS radio collars (fig. 1). Pride composition is well known for two prides and partially known for the third pride by the end of 2007. The Santawani pride (fig. 1, red and orange dots) consists of 7 adult females, 2 adult males 2 juveniles and 3 recently born cubs. The Gomoti pride (fig. 1, dark blue and light blue dots) consists of 4 adult females, 1 adult male, 3 juveniles and 4 cubs about six to nine months old. The Mogogelo pride (green dots) has 5 adult females, 2 adult males, 2 sub-adults and 7 cubs about six months old.

At this early stage in our lion population monitoring, the data currently in hand are not sufficient to provide reliable estimates of territory sizes and population density. This is due only to the short (three-month) period during which we have collected movement data and observed this subpopulation. Nevertheless, some noteworthy movements are apparent from the dataset to date

• The veterinary “buffalo” fence (in fig. 2) appears to represent an important boundary to the lions in the area, despite the fact we have observed it readily crossed by all carnivore species. Evidence from the recent ranging of the residents on the WMA side of the fence indicates they perceive the veterinary fence to be an important boundary. This either relates to learned risks associated with ranging on the livestock side of the fence, or the possibility that resident lions on the other side of the fence use it as a territory boundary that is observed by lions on both sides.
• Lions’ movement patterns seem to be highly influenced by the movements of conspecifics:
  • The female of the Mogogelo pride (fig. 2, green dots), together with another female and their 4 young cubs, seem to be confined within a 20 km. area, squeezed between the wide ranging of the Gomoti male to the west and north (fig. 2, dark blue dots) and the Santawani pride females to the east (fig. 2, orange dots). With the arrival of the rainy season at year’s end, the Santawani pride began moving east (fig. 2, orange arrows) perhaps releasing pressure on their western boundary.
  • Two weeks after professional hunters shot his coalition partner (possible brother), the “red” male (formerly the Santawani pride male) was pushed south (fig. 2, red arrow) by a new coalition of two males. Three weeks later he was found dead not more than 30m from the veterinary fence. The cause of death remains unknown following our investigation, but either poisoning or illegal shooting was considered likely.

Figure 2: GPS locations for 5 lions fitted with GPS radio collars in October 2007. A male and a female are collared in the Gomoti pride (dark blue respectively light blue dots). A female is collared in the Mogogelo pride (green dots). A female is collared in the Santawani pride (orange dots). A male from the Santawani pride (red dots) was found dead along the buffalo fence 3 weeks after being collared.

From the annual report submitted to the Government of Botswana by The Botswana Predator Conservation Program (Dr. J. W. “Tico” McNutt, Director)

Five adult leopards inhabiting the core study area that comprises the southern part of Moremi Game Reserve and the adjacent Wildlife Management Areas are currently regularly under observation for ranging, territoriality, and reproductive activity. Three males and one female are fitted with GPS radio collars and one female is fitted with a traditional VHF radio collar (fig. 3). A sixth uncollared adult female that holds a territory around the research camp has been identified and is opportunistically observed. A seventh territorial adult male, collared late in 2006 was shot in the study area in June 2007. By the end of October in the same year his territory (fig. 3, orange dots) was already being partially occupied by a young male (fig. 3, blue dots) estimated to be about 2 years old. The rapid immigration of a territorial male suggests that the sub-population of leopards in our study area is probably healthy and capable of rapidly recovering available habitat. However, the consequences of a human-caused removal (shooting) of territorial males on the subpopulation’s social structure and reproduction are as yet unknown. Mating has been observed repeatedly among the various resident males and females, but none of the closely observed females has been seen with cubs in the last 12 months. One unidentified adult female was reported to us with a cub in the study area north of the female (green dots, fig3)

Figure 3: GPS locations for 6 adult leopards inhabiting the core study area. Three males (yellow, red and blue dots) and two females (green and purple dots) are currently being monitored; professional trophy hunters shot the male (orange dots) in June.

Habitat use. As expected, the two sexes have completely overlapping territories. Whereas the two females have exclusive territories, the males show an unexpected high degree of overlap (fig. 3, yellow and red dots). Male territories do not typically overlap according to reports from other study areas. It may be that the “red” male is a young male looking for available habitat that is temporarily tolerated by the “yellow” territorial male. Estimates of territory sizes for the collared leopards are given in Table 3. As expected the territory size of males is significantly larger than for females, with a male roughly covering the territories of three resident females. The high variance within the territory sizes of the males is in part due to habitat differences. This density estimate is conservative since it does not take into account among same sex residents. More range data will be necessary to assess the degree of overlap between territorial individuals of the same sex and whether territorial residents of either sex commonly tolerate young individuals.

From the annual report submitted to the Government of Botswana by The Botswana Predator Conservation Program (Dr. J. W. “Tico” McNutt, Director)

In the last 12 months we were able to identify four adult female cheetahs that regularly roam through the study area that comprises Moremi Game Reserve and the adjacent Wildlife Management Areas. All females have been seen with cubs ranging between 3 and 10 months or in an advanced pregnancy status. This indicates that the local sub-population is healthy and successfully reproducing. Further research will allow us to assess the survival rate of the cubs. We were also able to identify two coalitions of two cheetah males each. Most likely due to the calving of the impalas and the resulting relatively high abundance of comparatively easy prey at year’s end, all cheetahs are well fed and in good conditions.

Figure 4: GPS locations for two male cheetahs over a period of 68 days (red dots) and 22 days (yellow dots). The “red” cheetah covers a territory of 280 km2. Both cheetahs show a preference for the open habitats (blue and green) associated with the Gomoti and Mogogelo rivers south of the “mopane tongue” (brown). The two adult females with VHF collars have extremely large ranges approximated by the shaded areas.

At present four of these identified individuals are collared. Two females are fitted with traditional VHF radio collars and two males (one in each coalition) are fitted with GPS radio collars. Both sexes range over large territories of several hundred kilometers. One of the two GPS collared males has a territory size of 280 kilometers (fig.2). The size of the territory is however likely to increase as more GPS fixes are collected. This translates in an overall low population density for the study area. More data are however required to assess the degree of overlap between individuals and accurately estimate the population density. The GPS data of both males also show a significant preference for the open habitats associated with the Gomoti and Mogogelo rivers south of the “mopane tongue” in the Moremi Reserve (fig. 2).

From the annual report submitted to the Government of Botswana by The Botswana Predator Conservation Program (Dr. J. W. “Tico” McNutt, Director)

Five packs of wild dogs resident in the study area in 2007 were closely monitored for ranging, habitat use, and reproduction. All five packs produced a litter of pups during the months June through August, the normal denning season. Lions disrupted the denning activities of the small pack of three adults in the Third Bridge area of Moremi Game Reserve (Table 4) and all the pups were subsequently lost. The other four packs in the study area successfully fledged pups from dens both in the Moremi and in the adjacent WMA’s, though lions have killed many of these pups in the last three months of the year. Several additional packs including a radio-collared pack on Chief’s Island also successfully produced pups elsewhere in Northern Botswana outside our study area. These included three packs known in the Linyanti/Selinda area, a pack in the Mapula/Vumbura area, and two known packs in eastern Ngamiland.

Table 4. Wild dog pack reproduction and pup survival through the end of 2007.

Most known mortality of pups to date is attributed to lion ambush and most of this occurs before the first half year of age. Using an average litter size of 9 pups per litter for the packs with unknown litter sizes, the survival of pups to 6 months of age in 2007 from this sample is 40 percent slightly below the 15-year average survival to one year (43%) from 71 litters. It is also worth noting the positive relationship in the data presented in Table 4 between pack size and pup survival a finding consistent with previous records for this population. Increased survivorship is presumed to relate to probability of ambush declining with increasing pack size due to increased vigilance.

From the annual report submitted to the Government of Botswana by The Botswana Predator Conservation Program (Dr. J. W. “Tico” McNutt, Director and Project lead: Tshepo 'Dungi' Kgokilwe")

Monitoring trends in animal population sizes and distribution is an essential component of wildlife research and management. The Botswana Predator Conservation Program is monitoring all five large carnivore species that occur in northern Botswana (Moremi Game Reserve and adjacent Wildlife Management Areas), including lions (Panthera leo), leopards (Panthera pardus), cheetahs (Acinonyx jubatus), spotted hyenas (Crocuta crocuta) and African wild dogs (Lycaon pictus). We are conducting spoor (tracks, trails and droppings) surveys with the aim to calibrate this indirect census method with the actual number of individuals per species identified from radio tracking data, in order to provide reliable and cost-effective indexes of large carnivore population sizes and distribution.

Figure 5. Overall spoor density (number of spoor per 100 km road) for all five large carnivore species occurring in northern Botswana (Moremi Game Reserve and adjacent Wildlife Management Areas).

A total of 14 different roads transects in three habitat types (mopane woodland, mixed woodland and floodplain habitat) are routinely surveyed for spoor. By September’s end, 110 transects, covering a distance of 1,897.3 km, have been driven. Spotted hyenas show by far the highest overall spoor density (number of spoor per 100 km road) in the sampling area, with cheetahs and wild dogs showing the lowest spoor density (Fig. 1). Lions, leopards and spotted hyenas show the highest spoor density in floodplain habitat, followed by mixed woodland and mopane woodland (Fig. 2). Cheetahs and wild dogs, on the other hand, seem to avoid floodplain habitat and show the highest spoor density in mixed woodland (Fig. 2).

Figure 6. Spoor density (number of spoor per 100 km road) for all five large carnivore species occurring in northern Botswana (Moremi Game Reserve and adjacent Wildlife Management Areas) according to habitat type.

As exemplified here by spotted hyenas as the most abundant large carnivore species that occurs in northern Botswana, we then relate these spoor density data to the actual number of individuals per species identified from radio tracking data. True density (number of individuals per 100 km) is estimated by dividing clan size by home range size (in km) times 100 (Table 1). This figure is regressed against spoor density within the respective clan’s home range (Table 1), each containing varying proportions of the three habitat types (Fig. 3).

Table 5. Characteristics of three neighbouring clans of spotted hyenas in northern Botswana (Moremi Game Reserve and adjacent Wildlife Management Areas).

*95% kernel of location data (n = 903, 3857 and 2261, respectively) from spotted hyenas fitted with GPS radio collars, with smoothing factor set at 0.5. Figure 7. Relationship between true density (number of individuals per 100 km) and spoor density (number of spoor per 100 km road) for spotted hyenas in northern Botswana (Moremi Game Reserve and adjacent Wildlife Management Areas).

Our spoor surveys thus provide a reliable and cost-effective indirect census method for monitoring trends in population sizes and distribution.

From the annual report submitted to the Government of Botswana by The Botswana Predator Conservation Program (Dr. J. W. “Tico” McNutt, Director and Project lead, PhD candidate: Hugh Webster)

Completing work from the 2005 and 2006 field seasons, I completed the sound playback experiments between lions (Panthera leo), spotted hyenas (Crocuta crocuta) and African wild dogs (Lycaon pictus). These playback experiments were designed to investigate in greater depth the relationship between these sympatric large carnivores in northern Botswana through measuring behavioural responses to interspecific vocalizations. I also completed the collection of the wild dog vocal lexicon and made an initial investigation (using playbacks) into the potential for wild dogs’ ultrasonic frequency calls to facilitate individual recognition.

This year was a productive final field season. I have been able to increase the sample sizes for various playback experiments, increasing the strength of significance of the previously noted reactions. Wild dogs flee from lion roars but not spotted hyena whoops; lions approach wild dog vocalizations, but ignore control playbacks of birdcalls; and spotted hyenas either approach or ignore wild dog vocalizations. In addition, I have been able to look more closely at trends suggested by early analyses, such as the affect on response time related to differences in habitat. In comparatively dense vegetation, for example, wild dog visibility is impaired and risk of ambush is predicted to increase. Accordingly, wild dogs were observed to flee more quickly from lion roars in relatively closed habitat, lending reliability to the assumption that behavioural responses reflect perceptions of risk.

I also finished with playback experiments designed to compare the reactions of spotted hyenas versus wild dogs to lion roars. Preliminary results are interesting and suggest that, although spotted hyenas are also inclined to move away from lion roars, they move significantly shorter distances than wild dogs.

During the wild dog denning season (late May to mid September), I recorded both adult and pup vocalizations, providing more examples of types of calls adding to those catalogued to date. This set of recorded calls will allow discriminant function analysis to be used in defining the parameters of the wild dog calls, and will move the wild dog vocal lexicon from previous largely subjective definitions to a statistically based set of definitions. Furthermore, the extent to which wild dogs utilize ultrasonic frequencies in their calls was recorded for the first time.

My initial hypothesis was that wild dogs produce ultrasonic frequency calls to allow communication with the same species, while avoiding the potential costs (kleptoparasitism and predation) incurred by attracting eavesdroppers (lions and spotted hyenas). However, it became clear after several recordings that on the vast majority of occasions, wild dogs produce normally audible and often relatively loud calls concurrently with the ultrasonic frequencies. Therefore I designed a set of playback experiments that would test the wild dogs’ ability to discriminate between familiar (own pack dominant female begging) and unfamiliar (unrelated pack dominant female begging) calls (as an alternative explanation for the function of these ultrasonic calls) with and without the high frequency element included (Fig. 8). The results suggest that wild dogs are able to discriminate between calls regardless of whether the high frequency element is included (response to unfamiliar calls was stronger for 12 out of 14 pairs), but there is an intriguing trend suggesting that they may be able to do so better with the high frequency element present (response to unfamiliar calls with high frequency element included was stronger for 5 out of 7 sets of calls).

Figure 8. Begging calls of a dominant female wild dog with the high frequency element left in (a) and filtered out (b).