Cheetah (Acinonyx jubatus)- Data, Pictures & Videos

[Acinonyx sp.]

10.1 INTRODUCTION 
Extensive information regarding the feeding ecology of cheetahs has been collected from the Serengeti (Caro 1994, Frame 1986, Kruuk and Turner 1967, Schaller 1968), where 21 prey species were recorded, ranging in size from mole rats (Cryptomys spp.) to wildebeest, with a strong bias towards Thomson’s gazelles. Other studies in East Africa (Burney 1980, Eaton 1974, Graham 1966, McLaughlin 1970) have also revealed preferences for gazelles (Gazella spp.) and impala, amongst a diverse prey base. In northern Kenya, cheetahs were observed taking kudu, gerenuk and dik-dik (Hamilton 1986) while kob and oribi have been noted as prey in west Africa (Nowell and Jackson 1996). Data from Kafue National Park, Zambia showed puku to be the favoured prey species (Mitchell et al. 1965), while cheetahs in the Lowveld Region of South Africa (Hirst 1969, Pienaar 1969) took a preponderance of impala amongst 15 species taken. In the Southern Kalahari, (Mills 1984) found that cheetahs killed prey ranging from bateared foxes (Otocyon megalotis) to wildebeest, with springbok as the favoured species. The summary to date, then, is that cheetahs predominantly kill medium-sized (10 – 35 kg) antelope, but will opportunistically take other prey if available. Against this background, the diet of cheetahs on Namibian farmlands is interesting for two reasons. First, the cheetahs in this habitat exist in a highly managed ecosystem, where kleptoparasites such as spotted hyaenas and lions have been eliminated, whereas previous studies of cheetah diet have been conducted in areas where they are sympatric with other larger carnivores, by which they are potentially disadvantaged by intra-guild competition. It is interesting to investigate how their dietary preferences change in the absence of such competition. Secondly, this population is threatened by farmers who kill them because of he perception that cheetahs kill substantial numbers of domestic stock and ranched game, particularly expensive and exotic game. Cheetahs are known to kill smallstock and calves up to six months old (Marker-Kraus et al. 1996), but it is important to investigate whether the level of predation corroborates the perception of them being a serious problem.

Picture 10.1 Calf injured by suspected cheetah attack – such incidents can cause significant economic losses to farmers. 

*This image is copyright of its original author

Through surveys conducted with Namibian farmers, the cheetah was reported to prey upon a wide range of species on the farmlands, including livestock as well as both indigenous and exotic game species (Marker-Kraus et al. 1996). This paper aims to identify the relative importance of the different species in the diet of cheetahs on Namibian farmlands, so that problems and potential solutions can be identified in order to develop suitable cheetah conservation strategies.

Diet estimation of carnivorous mammals can be tackled by various methods, each subject to different biases (Mills 1984, Reynolds and Aebischer 1991). Opportunistic and direct observation of kills, while the predominant method for large carnivores in East and South Africa, is impractical in the dense bushveld of Namibian farmland. The traditional solution involves quantification of undigested prey remains in scats (e.g. Emmons 1987, Geffen et al. 1992, Lanszki et al. 1999, Previtali et al. 1998). However, it has long been obvious that extrapolation from volumetric analysis of undigested prey remains in faeces is an unsafe basis for quantifying carnivore diet unless corrected for differential digestibility of different prey sizes and species (Ackerman et al. 1984, Floyd et al. 1978, Lockie 1959, Scott 1941). Such uncorrected extrapolation risks, for small prey, the overestimation of biomass and under-estimation of numbers consumed. Therefore, as one step in our diet analysis we calculated digestibility indices for captive cheetahs following the protocol established by (Floyd et al. 1978) for grey wolves (Canis lupus), and used these indices to estimate rates of livestock depredation caused by cheetahs on the farmlands. We also compared estimates of cheetah diet derived by contrasting methodologies (e.g. faecal analysis versus aerial surveys of kills) to evaluate the biases inherent in each.


10.2 METHODS 
10.2.1 Feeding trials 
Following (Floyd et al. 1978), nine trials were conducted in two 256m2 captive holding pens at the Cheetah Conservation Fund’s research farm near Otjiwarongo, Namibia. Before each trial, the cheetahs were fasted until no fresh scats were being produced, a process which took 40-96 hours. This was similar to fasting periods experienced in the wild: Caro (1994) reported fasting times of 30-36 hours, McLaughlin (1970) reported fasts of 48-72 hours, and Broomhall (2001) described fasting periods ranging from 84 to 168 hours. Carcasses were weighed and then fed intact to the cheetahs. Five carcasses were fed to two wild-born, two-year old females. Four carcasses were fed to three wild-born, three-year-old males. Four species were used with prey weights <30kg (hare, lamb, goat and steenbok); while two species were heavier, namely kudu and oryx. Since a high percentage of cheetah kills are either abandoned after gorging or are stolen by a competing predator (Caro 1994), the carcasses were removed when all feeding cheetahs remained lying down for more than 10 minutes without returning to feed (33-125 minutes). After feeding, the carcass was removed and weighed to the nearest 0.5 kg. Scats were collected twice daily, in order to minimise both trampling and desiccation. Scat consistency varied from liquid or semi-liquid scats that would be unlikely to be found and collected during a field study, which were categorised as noncollectable scats (NC), to firmer scats that were likely to be found and collected in the field (field-collectable, FC). Field-collectable scats were counted and weighed immediately after collection. Statistical analyses were conducted using SPSS version 10.05 (SPSS, Chicago, Illinois). Kolmogorov-Smirnov and Shapiro-Wilk tests were used to investigate normality, and non-parametric procedures were used where there was significant deviation from normality. Analysis followed (Floyd et al. 1978) and (Weaver 1993), using a least squares regression plot, which yielded a regression equation, where y is the kg of prey consumed per collectable scat and x is the average weight of an individual of a given prey species. By multiplying y by the frequency of occurrence (n) of each prey species in the sample, it was possible to obtain a total weight consumed of each species and calculate the ratios of biomass consumed between prey species. The total weight of each species consumed was then divided by the average estimated weight in order to compute the number of individuals consumed, and ratios were computed relative to a kudu calf weighing 16kg. Masses of subadult animals were used for eland, oryx, kudu and red hartebeest, as cheetahs most commonly prey upon the calves of these large species rather than hunting adult animals (Marker-Kraus et al. 1996). 

10.2.2 Scat analysis 
Scats were collected from wild cheetahs that were live-trapped by farmers (as described in Chapter 3), both from the traps themselves and during examination, and were also collected in the field, particularly from ‘playtrees’, which are certain trees used by cheetahs for scent-marking with urine and faeces (Marker-Kraus and Kraus 1995). Once collected, scats were individually placed in nylon stockings and washed through two complete regular cycles in a conventional washing machine. No bleach or detergents were used. The washing process left in the stocking only hair, bones, teeth and hooves, and the stockings and their contents were then hung out to dry. The dried remains were spread evenly into a dissecting pan with a grid base of six 67.5cm2 squares, and one hair was randomly sampled from each square, carefully examined, and cuticle scale imprints were made. Hairs were sandwiched between two glass slides on a plastic cover slip, held together by four small (no. 20) binder clips, and heated for five minutes in a toaster oven at 108°C, removed, and left to air cool. The hair was then gently removed from the cover slip using forceps or fingernail, and the hair’s scale characteristics were used to determine which species it was from. Macroscopic distinctions narrowed the options and cuticle imprints finalized the identification. Kudu and eland hairs were often difficult to distinguish, so were categorized together in some instances. In compiling our reference collection, we were mindful of Koegh’s (1983) result that hair from fresh carcasses and preserved skins is identical (Buys and Koegh 1984, Koegh 1983). Our collection involved hairs and imprints from neck, back, belly, and hindquarter regions of each possible prey species in the study area.10.2.3 Information on kills from radio-tracking flights and from farmers Between 1993 and 1999, radio-collared cheetahs were tracked on a weekly basis from a fixed-wing Cessna 172, as described in Chapter 11. During these flights, cheetahs were occasionally sighted on identifiable kills. Although they may do so exceptionally (Caro 1982, Pienaar 1969, Stander 1990), cheetahs do not generally scavenge from other predators (Caro 1994, Wrogemann 1975) and we therefore assumed that the cheetah had killed the animal being eaten. We also recorded whether the cheetah was sighted within 500m of livestock or game. Interpretation of the scat analysis data was also made in the context of farmers’ answers during a questionnaire survey conducted between 1991 and 1999 regarding their observations and perceptions of cheetah predation (Marker-Kraus et al. 1996), as detailed in Chapter 12. The results of the feeding trials and corrected scat analysis were used to estimate rates of livestock depredation by cheetahs in the study area.

10.3 RESULTS 
10.3.1. Feeding trials S
cats containing the presented prey item were produced from 48 to 111 hours after feeding, and the feeding trial results are shown in Table 10.1.
Table 10.1 Results of feeding trials performed on captive Namibian cheetahs. Scats were classified as either field-collectable (FC) or non-collectable (NC). 

*This image is copyright of its original author

Of the four smaller prey species, the mean percentage consumed was 69.7%, but for the two species of large antelope this dropped to 16.8%. There was a strong correlation both between prey mass presented and prey mass consumed (rs = 0.86, p = 0.007, n = 8), and between prey mass presented and fresh collectable scat weight (rs = 0.74, p = 0.038, n = 8). However, the smaller prey items consumed gave a proportionally greater fresh weight of field-collectable scats in relation to the prey mass presented, with the mass of field-collectable scats averaging 8.1% of the prey mass presented for the four smaller species, but only 1.2% for the kudu and oryx. The number of field-collectable scats per kg of food consumed diminished with increased prey weight – on average, the four small species gave 2.4 field-collectable scats per kg of prey eaten, while kudu and oryx gave a mean of 0.8 scats/kg. Data summarised in Table 10.2 revealed a strong correlation (r = 0.89, p = 0.017, n = 6) between the weight of prey consumed per collectable scat and average weight of the prey species presented.

Table 10.2 Summary of results from the feeding trials for each prey species presented. 

*This image is copyright of its original author

A regression on these variables generated the following equation: y = 0.0098x + 0.3425, which can be used to provide valuable information on the relative contribution of different prey species reported as part of the cheetah’s diet (Marker-Kraus et al. 1996). This information is shown in Table 10.3. 

Table 10.3 Ratios of prey animals consumed, using the corrected scat analysis, for 100 scats containing prey species.

*This image is copyright of its original author

*This image is copyright of its original author

10.3.2 Scat analysis 
Ninety-eight cheetah scats were analysed, of which 79.6% (n = 78) were from cheetahs held for 4 days or less, and 20.4% (n = 20) were from cheetahs held captive for over 4 days. From the feeding trial results, only cheetahs that had been captive for 4 days or less were considered to be indicative of diet in the wild, as any scats produced after this time would not reflect diet before capture. Of the 78 scats from wild cheetahs, 33.3% (n = 26) were from game farms (either from captured cheetahs or collected from playtrees), 48.7% (n = 38) were from livestock farms, and 17.9% (n = 14) from unknown locations. Table 10.4 presents the total number of scats collected from wild cheetahs, location of collection and the prey species identified in them. In the majority of cases, the cheetahs appeared to be predating upon indigenous game species, while in 6.4% of cases the prey species identified were domestic stock. 

Table 10.4 Contents of wild cheetah scats collected from various locations on the Namibian farmlands. 

*This image is copyright of its original author

*This image is copyright of its original author

Applying corrections for differential digestibility (Table 10.5), the prey selection can be more accurately determined. Only the scats where kudu and eland hairs could be distinguished were used for these two species. 

Table 10.5 Ratios of prey animals consumed, calculated using the corrected scat analysis.

*This image is copyright of its original author

This table highlights the importance of applying correction factors to scat analysis to avoid under-representing the consumption of smaller prey animals. For instance, although hare remains were found in only 3 scats and accounted for only one fifth of the biomass that kudu did, we deduced that nearly twice as many hares as kudu were preyed upon. Conversely, a similar biomass of eland and kudu appeared to have been consumed, but use of the correction factor indicated that fewer than half the number of eland would have been killed compared to kudu. Forty-six scats were analysed from wild cheetahs of known sex (37 from males and 9 from females), and identifiable prey remains were found in 27 cases, from 23 male and 4 female cheetahs. A higher percentage of scats from male cheetahs contained the remains of large antelope species (kudu, eland, red hartebeest and oryx), while those from females more frequently contained evidence of smaller antelope such as steenbok (Figure 10.1). The remains of domestic stock were found only in scats collected from male cheetahs, but the sample size of scats from female cheetahs was too low to draw any substantial conclusions from this. 

Figure 10.1 The percentage of scats from male and female cheetahs that contained remains of large and small species of antelope, other species such as hares and birds, and domestic stock.

*This image is copyright of its original author

10.3.3 Additional information regarding kills 
Between 1993 and 1999, 325 visual observations of radio-collared cheetahs were made. From these observations, 21 cases were recorded of cheetahs on identifiable kills, and the prey consumption determined using this method was compared to that from the corrected scat analysis (Figure 10.2).

*This image is copyright of its original author

Figure 10.2 Estimates of the relative contributions of different prey species to the diet of Namibian cheetahs, using data from observed kills during radio-tracking flights, and from corrected scat analysis. 


Even when limited to the prey species identified through both techniques, the composition of the diet indicated by visual observations and corrected scat analysis differed significantly (Ȥ 2 = 33.1, df = 4, p = 0.000). Aerial sightings led to higher representation of kudu, red hartebeest and oryx than estimated through the scat analysis, with the other species less represented. In the instances where cheetahs were located near potential prey (n = 1088 locations), they were within 500m of game species 77.6% of the time and to livestock 22.4% of the time. From the farm survey conducted, 58.6% (n = 81) of the farmers believed that kudu calves were the primary prey of the cheetah. Springbok, where regionally available, was also reported as a main component of the diet, as were warthog and steenbok. Oryx and hartebeest calves were considered to be common prey, followed by a variety of other animals including duiker, eland, ostrich, small game birds, guinea fowl (Numida meleagris), kori bustards (Ardeotis kori), and hares

10.3.4 Estimating rate of livestock depredation caused by cheetahs The maximum rate of food consumption for wild cheetahs has been estimated as 5.5kg/cheetah/day (Eaton 1974), which equates to 1958 kg of prey consumed/cheetah/year. Our feeding trials revealed that 1.87 field collectable scats were likely to be produced per kilogram of prey consumed, leading to an estimated production of 3661 field-collectable scats/cheetah/year. The scat analysis showed that on the Namibian farmlands, 4.3% of scats collected contained evidence of domestic calf consumption, while 2.1% contained sheep remains. Using the correction factor, we calculated a consumption of 0.018 calves for each scat containing calf remains, while the figure was 0.016 individuals for sheep. Therefore, out of 3661 scats, 157 would be likely to contain calf remains and 77 would contain sheep remains, indicating the consumption of 2.8 calves and 1.2 sheep per cheetah per year. However, Schaller (1972) calculated that cheetahs killed 35% more prey than they consumed, leading to estimated kill rates of 3.8 calves and 1.6 sheep per cheetah per year. Assuming a minimum density of 2.5 cheetahs per 1000km2 on the farmlands and an average farm size of 8000ha (80km2 ) (see Chapter 11), the minimum rate of livestock depredation due to cheetahs can be calculated as 0.01 calves and 0.004 sheep per km2 , or 0.76 calves and 0.32 sheep annually on an averagely-sized farm. The accuracy of these calculations obviously depends on the density of cheetahs living in the study area, estimates of which vary widely (Stander 2001). Using the highest reported estimates of cheetah density on the Namibian farmlands (34 cheetahs/1000km2 : Stander 2001), the approximate rates of livestock depredation due to cheetahs would be 10.3 calves and 4.4 sheep per farm per year. 10.4 DISCUSSION The feeding trials supported the conclusions of Lockie (1959) and Floyd et al. (1978), that, if analysis is based on uncorrected volumetric measures of undigested remains in scats, then smaller prey items are over-represented in biomass consumed but under-represented in numbers if uncorrected scat analysis is performed. The cheetah is an opportunistic predator whose prey varies in size from rodents to adult ungulates (Burney 1980, Caro 1994, Frame 1986, Marker-Kraus et al. 1996, Schaller 1968), and this great variation in prey size makes accurate interpretation of scat analysis more complicated. Consumption of smaller prey gave a higher number of field-collectable scats relative to the mass consumed, because they are composed of relatively more indigestible matter. Feeding on meat alone, rather than bone and hide, tends to result in the production of more liquid scats, and these would probably not be collected during field studies (Ackerman et al. 1984, Floyd et al. 1978). This is likely to be of particular importance regarding cheetah dietary habits due to their method of prey consumption. Although cheetahs are known to consume some bone (Phillips 1993), they consume more pure muscle (rather than skin or bone) than do other large carnivores (van Valkenburgh 1996, Wrogemann 1975), and this is likely to be even more pronounced when eating from a large carcass. Use of correction factors is therefore very important for accurately estimating cheetah diets.Accurate analysis of wild cheetah diet relies upon the collection of enough scats from which prey remains can be identified. Using the equations in Reynolds and Aebischer (1991), 9600 scats containing identifiable prey remains would be required to establish that these estimated prey proportions are accurate. Given that only 76.9% of the wild cheetah scats analysed contained identifiable prey remains, it would necessitate 12 500 scats to achieve the aforementioned statistical power, which is unrealistic. Our experience has shown that collecting cheetah scats is very difficult due to several factors, including the large home ranges cheetahs occupy (Marker 2000) and the rapid desiccation of scats in arid environments. Additionally, scats are hard to collect from cheetahs trapped by farmers, as the cats have often gone without food for several days and any scats in the traps are frequently trampled. The available data, therefore, based on a much smaller sample size, can only give a basic insight into the dietary habits of cheetahs on Namibian farmlands. Collecting scats from playtrees and trapped cheetahs biases the data towards males, as the majority of cheetahs visiting playtrees and being trapped are male (Marker-Kraus et al. 1996). Male cheetahs are likely to take larger prey than do females (Mills 1992), so the prey selection determined during this study may not be entirely representative of female cheetahs. Additionally, the interpretation of the scat analysis in terms of numbers of prey animals consumed assumes that the prey animals taken weighed approximately the average masses shown. However, despite these limitations, and especially given the lack of other information available, these data can contribute usefully to understanding the diet of wild Namibian cheetahs on farmlands. The radio-tracking data reveal that cheetahs are sighted near livestock relatively frequently, and this is exacerbated by the species’ diurnal nature and consequently greater visibility than other predators. Such sightings by farmers who are experiencing stock loss potentially leads to the assumption that cheetahs are the cause, and creates the perception of them as being frequent stock-killers. The corrected scat analysis indicates, however, that cheetahs preferentially take wild game species over domestic ones. Although 38 scats were collected on livestock farms (over half from cheetahs that had been trapped as a supposed threat to livestock), only 2 of those contained any evidence of domestic stock consumption. The fact that domestic stock was evident in 6.4% of the scats does verify that cheetahs prey upon livestock, but as two-thirds of the available prey base is livestock (Marker-Kraus et al., 1996), cheetahs appear to show selection towards game species. It is difficult to estimate rates of livestock depredation due to cheetahs from this information, as estimated figures for cheetah density in the study area vary greatly (Stander 2001). In a recent survey (see Chapter 12) farmers in the region reported losing an average of 0.9 calves and 1.3 smallstock annually to cheetahs, which was slightly higher than estimated using the minimum density figures, but far less than would be expected if cheetahs exist at maximum density. Conducting further research into gaining a more accurate estimate of cheetah density will be vital for independently examining the level of stock loss that cheetahs are likely to be responsible for. Relying therefore on reports by farmers (see Chapter 12), the level of livestock depredation attributed to cheetahs was substantially less than that caused by other predators, and indicate that livestock depredation due to cheetahs is unlikely to be a major financial burden for Namibia’s commercial farmers. However, the predominance of game species in the diet does mean that the cheetah is likely to be perceived as a threat on game farms. Many game farmers stock exotic game species on their land for trophy hunters, which are more valuable economically than indigenous game but can be more liable to predation than the better adapted indigenous species (Marker and Schumann 1998, Marker-Kraus et al. 1996). Although these results suggest that cheetahs are preying mainly upon indigenous game species rather than the more expensive exotic game, losses to large carnivores remain a potential problem for game farmers. 

*This image is copyright of its original author

Picture 10.2 Introducing game species into areas where they would not naturally occur can have serious consequences in terms of losses. Species such as springbok, which are adapted for short-grass plains, are highly susceptible to predation in the thickly bushed farmland habitat, and often suffer high levels of depredation, reducing game farmers’ tolerance for predators. 

In line with comparable studies of other carnivores, (e.g. Karanth and Sunquist 1995, Mills 1992) the diet estimated from sighted kills contained a greater proportion of large prey than did that estimated from faeces. The only exception in this study was for eland, where fewer kills were seen than would be expected from the scat analysis. This may be due to the fact that eland are nomadic (Smithers 1983) and for much of the time would not be on the relatively small area of farmland where radio-tracked cheetahs were being followed. The wild prey base available to the cheetah is critical in the issue of predator conflict. According to many Namibian farmers, maintaining a higher ratio of wildlife to cattle is the most important feature in reducing livestock predation in the survey area (Marker-Kraus et al. 1996), as a plentiful wildlife population provides an abundance of prey, which in turn reduces the farmers’ conflict with predators. Overall, these data indicate that cheetahs show selection for indigenous game species rather than for domestic stock. However, even a relatively low level of predation upon expensive, introduced game, or upon livestock, can have a serious economic impact on farmers (Oli et al. 1994). In order to conserve cheetahs successfully on farmlands and reduce the level of removal, strategies must be found that mitigate the economic loss caused by them. Fenced sections of farms containing expensive game animals can be protected through effective maintenance of perimeter fences, or, more sustainably, by the removal of game fencing and the development instead of cooperative game management areas in the form of conservancies.

*This image is copyright of its original author

Picture 10.3 Forming conservancies, where game is managed collectively rather than being split into small, game-fenced areas, provides several advantages such as allowing the natural movement of game populations across the farmlands, and reducing the impact of losses due to predators for individual farmers.

There are also several livestock management practices, such as the use of guarding animals, calving corrals and synchronized breeding seasons, that have been shown to be effective in reducing stock losses both to cheetahs and other predators (Marker-Kraus et al. 1996). Additionally, the development of ecotourism and sustainable trophy hunting both have the potential to turn predators into an economic asset rather than a detriment to the farmers on whose lands they survive.

http://citeseerx.ist.psu.edu/viewdoc/dow...1&type=pdf Page:214-233 Reply