The leopard, Panthera pardus, is particularly threatened outside conservation areas in South Africa. This has been attributed to a reduction in natural habitat, decreasing natural prey populations, and commercial exploitation such as trophy hunting, and most importantly, persecution by stock farmers (Woodroffe 2001). The leopard population in the Baviaanskloof Mega-Reserve (BMR) has undergone a substantial decrease in range and numbers in the past 200 years, resulting in a highly fragmented population in the Baviaanskloof Mega-Reserve, and is regarded as being insecure. There is thus a need to investigate the nature and extent of leopard-stock farmer interactions to provide the foundation for an effective leopard conservation plan. Here I investigated the ecological and socio-economic factors influencing leopard-stock farmer conflict via landowner surveys and estimated potential leopard numbers using a prey-based density model. Leopards are not necessarily the most important causes of livestock mortality in the BMR. On average, leopards killed significantly less livestock (0.7 percent livestock per year) than black-backed jackals (4.7 percent per year) and caracal (2.5 percent per year), yet 67 percent of farmers had negative attitudes towards leopards. These negative attitudes were not significantly related to stock losses. However, most of the farmers that had negative attitudes towards leopards did not have any stock losses attributed to leopards. Thus if predator-stock conflict is not reduced it will result in the retaliatory killing of leopards. This will have severe consequences for this relatively small population (estimated at 59 – 104 individuals by the prey-based model), which may ultimately lead to the local extinction of these leopards (Woodroffe & Ginsberg 1998).

The trophy hunting of African leopards Panthera pardus pardus may generate revenue to help foster their conservation. However, leopards are sensitive to hunting and populations decline if overharvested. The practice therefore requires careful management grounded in robust estimates of population density/status. Camera-trap surveys are commonly used to establish leopard numbers, and may guide harvest quotas. However, such surveys are limited over wide spatial scales and many African governments lack resources to implement them. In this thesis I explore the potential use of a harvest composition scheme applied to puma Puma concolor in North America, to monitor leopards. The method hinges on the susceptibility of different leopard cohorts to hunting and if this varies, then predictions can be made about harvest composition. Susceptibility is likely to be governed by space use, encounter rates with bait lures (a common method used to attract leopards to hunting hides) and hunter selectivity. Thus in this thesis I explore leopard susceptibility to these factors using a protected leopard population in northern Zululand, South Africa. In my first chapter I examine using scent lures in camera-trapping. Against a backdrop of a passive survey I show adult males, females and sub-adults are captured at similar rates compared to a passive survey using lures. The use of lures does not appear to violate closure assumptions or affect spatio-temporal patterning, but their use appears limited as density estimate precision is not improved. My second chapter examines ecological (likelihood of encountering a hunter) and anthropogenic (attractiveness to hunters) susceptibility of leopards to trophy hunting. I show that adult males are the most susceptible cohort to hunting (sub-adults least susceptible). I then take the incident rates from ecological and anthropogenic models and create a theoretical harvest composition using population parameters of protected leopards. My third data chapter departs from hunting susceptibility and examines determinants of leopard trophy package price across Africa. I show that factors such as trophy quality, outfitter leopard hunting reputation and hunt success have little impact on price determination. Instead, overall outfitter reputation and the number of charismatic species in a package are positively correlated with price. These results have important consequences on several sustainable leopard hunting schemes proposed in the literature.

Trophy hunting is often employed as a conservation management tool for large predators. However, in order for this method to succeed, hunting levels must be sustainable. Very little robust population data exist for African leopards (Panthera pardus) in general, and almost no density or spatial ecology data exist for leopards in Zimbabwe. Zimbabwe has one of the highest annual CITES leopard trophy hunting quotas in Africa, the sustainability of which has not been assessed, despite large scale landuse changes over the last 12 years. The focal area of this study was within the Mangwe district, in the south-west of Zimbabwe. The region is dominated by cattle and wildlife ranches, with high levels of leopard hunting, making it an important area for assessing leopard population density and spatial ecology. Three population density estimation methods were employed in my study: a spoor index survey, an unbaited camera-trapping survey and a baited camera-trapping survey. Using three calibration equations, spoor indices appeared to underestimate the leopard population (1.28-3.29 leopards/ 100 km²) as the equations were calibrated for areas with different habitats and leopard densities. In addition, the unbaited camera survey only produced six leopard photographs, unsuitable for individual identification and analysis. By contrast, the baited camera survey produced 292 identifiable leopard photographs, from which 13 individuals were identified. Density estimates calculated using the programme CAPTURE and the M[subscript]h model with the Mean Maximum Distance Moved Outside of Study Area (MMDMOSA) buffer method (4.79±0.83 leopards/100 km²), and the programme SPACECAP, using a buffer of 2.5 km (5.12±0.62 leopards/100 km²), appeared to generate the most reliable leopard population estimates. To assess the spatial ecology, three leopards (one male, two females) were captured and fitted with GPS collars. The home range estimates of the three leopards (95% Kernel UD: male 263 km², females 31 and 45 km²) were smaller than those of leopards in more arid regions, but larger than those of mesic habitats. This suggests that the Mangwe area has a higher quality habitat than the arid regions of Namibia, but less suitable habitat than protected bushveld areas (e.g. Kruger National Park, South Africa). My data represents the first robust leopard density and home range assessment for Zimbabwe. In addition, my results indicate that the current hunting quota issued to the Mangwe area is unsustainable. Consequently, I recommend revising the quota to five leopards for the entire area, and halving the current national leopard quota to 250, until a national leopard census is completed.

Leopards (Panthera pardus) are the most common large predators, free roaming outside of protected areas across most of South Africa. Leopard persistence is attributed to their tolerance of rugged terrain that is subject to less development pressure, as well as their cryptic behaviour. Nevertheless, existing leopard populations are threatened indirectly by ongoing transformation of natural habitat and directly through hunting and conflict with livestock farmers. Together these threats may further isolate leopards to fragmented areas of core natural habitat. I studied leopard habitat preferences, population density, diet and the attitudes of landowners towards leopards in the Little Karoo, Western Cape, South Africa, an area of mixed land-use that contains elements of three overlapping global biodiversity hotspots. Data were gathered between 2010 and 2012 using camera traps set up at 141 sites over an area of ~3100km², GPS tracking collars fitted to three male leopards, scat samples (n=76), interviews with landowners (n=53) analysed in combination with geographical information system (GIS) layers. My results reveal that leopards preferred rugged, mountainous terrain of intermediate elevation, avoiding low-lying, open areas where human disturbance was generally greater. Despite relatively un-fragmented habitat within my study area, the leopard population density (0.75 leopards/100km²) was one of the lowest yet recorded in South Africa. This may reflect low prey densities in mountain refuges in addition to historical human persecution in the area. Currently local landowners are more tolerant of leopards than other wildlife species with incidents of conflict involving leopards being rare relative to black-backed jackals (Canis mesomelas), baboons (Papio hamadryas), caracals (Caracal caracal) and porcupine (Hystrix africaeaustralis). Although current levels of conflict between leopards and stock farmers are low, leopards do depredate livestock, which constitute 10-15% of their diet. Improved livestock husbandry measures and co-operation between conservation authorities and farmers are necessary to mitigate such conflict and balance economic security with biodiversity conservation in the region. Leopards are the only remaining top predators throughout much of the Little Karoo and the Western Cape and as such are predicted to play a critical role in ecosystem structure and the survival of other species. Current high levels of connectivity between areas of suitable leopard habitat bode well for the conservation status of leopards within this region and future conservation efforts need to ensure that narrow corridors linking such habitat are preserved. The potential for leopards to serve as both an umbrella and a flagship species for biodiversity conservation suggests that long term monitoring of this population would be a conservation priority for the Little Karoo.

The conservation of leopards in the Eastern Cape Province requires a holistic approach that considers both predator-human interactions as well as the biology of the carnivore. Numerous studies have been conducted on leopards within protected areas; however more information regarding the species is needed outside these areas to facilitate effective management of predators. The spatial ecology of the leopard (Panthera pardus) were studied in the Baviaanskloof and GAENP in the Eastern Cape. The Baviaanskloof is an extensive area of mountainous terrain (approximately 2665km²) which has a mosaic of land uses, and leopards move from conservation areas to farmland where they come into contact and conflict with farmers. This study examined the spatial ecology of leopards living on farmlands adjacent to protected land. The space utilization and activity patterns of six leopards were analysed. These animals were caught and released on farmlands in the Baviaanskloof (n=4) or translocated (n=2) when not possible to release on site. The animals were caught by means of fall-door, walk-in traps and fitted with Vectronic GPS collars that facilitated the collection of high quality GPS fixes from each animal. Data was collected using VHF and UHF telemetry to download data. An understanding of spatial requirements in areas with different land use, and the extent of overlap of space use with other leopards allow, for the first time in the region, the calculation of possible maximum population size. Analysis of range size was carried out using two methods: minimum convex polygon, and Kernel Utilization Distribution. Finally, a key predictor of space use is prey availability. I assessed the prey base using a grid of camera traps. The studied leopards revealed large range utilization with minimal overlap. The activity patterns suggest there is no preference between diurnal and nocturnal activity patterns and the population density was estimated between 0.3 - 1.3 leopards per 100km². Large home range sizes and low population densities suggest that leopards require large areas of suitable habitat, and that conservation efforts need to be extended beyond protected areas to ensure the long-term viability of leopard populations in such areas.