The elephant population of South Africa's Addo Elephant National Park increased in number from 11 elephants when the park was created in 1931, to 284 elephants in 1998. We studied management records throughout this period in conjunction with demographic data, enabling retrospective assessment of the long-term impacts of management actions. Problem animal control during the 1930s left no sexually mature bulls in the population for a period of 9 years, hindering the population's initial recovery. Population growth prior to 1954 was also limited by high mortality: between 1931 and 1954 inadequate fencing allowed elephants to stray outside their protected area resulting in elephants being shot by farmers or dying from collisions with trains. Secure elephant-proof fencing was constructed in 1954. Subsequently, there was a significant decrease in mortality (from 5.0% to 1.2%) and an increase in population growth (from 3.2% to 6.1%). However, evidence suggests that confinement might have had a negative impact on survival and social behaviour of adult males. We suggest that the concepts of social carrying capacities and behavioural conservation of populations need to be incorporated into management. The case studies described here highlight the importance of treating conservation management and research as necessary partners, and monitoring should be an integral part of any management plan.

The majority (98%) of the 174 female elephants in South Africa's Addo Elephant National Park are tuskless. Long-term records are used to assess trends in tusklessness in this elephant population throughout the 20th century. These data are then used to test hypotheses of possible causes of this tusklessness. It is shown that attributes of the region's vegetation (structure and/or nutrient composition) and selective hunting cannot provide adequate explanations for the high frequency of tusklessness among the Addo elephants. Non-selective genetic changes resulting from the population's isolation, small size and bottlenecks (during the 1800s and 1920s) are suggested to be primarily responsible for the observed tusklessness. The rapid increase in the frequency of tusklessness during the past 70 years provides a phenotypic indicator of underlying genetic drift.