USING ELEPHANT CARCASS RATIOS TO DETERMINE POPULATION TRENDS

Summary. Many African elephant populations have declined over the last two decades (Douglas-Hamilton. 1987). In most census zones, as the number of elephants decreases, the number of dead elephants increases. By counting both live and dead elephants a carcass ratio can be derived. This is the proportion of dead elephants to all elephants dead and live, and has been used as an index of relative elephant mortality (Douglas-Hamiton and Hillman, 1980). In this paper successive counts of dead and live elephants in central, east, and southern Africa have been used to plot elephant trends against carcass ratios. It was found that the carcass ratio was correlated with the rate of decrease over a 4-yr period. The model was applied to new census data from Selous and Kilombero and found to give a close prediction of actual trends. The model was then tentatively applied to census zones in Tanzania where only a single count had been made. It suggested that 2 of the 16 regions surveyed had stable elephant populations, but that several areas (such as Maasai Steppe and Tabora) were undergoing rapid declines in elephant numbers. The model suggested that Serengeti in 1977 had probably already suffered a 21 % decline in the previous 4 yr, but that the decline outside the park had been more severe at 57%. Ruaha national park in 1977 it suggested had increased by 7 % over the previous 4 yr, but had decreased by 9 % in the surrounding areas. Selous game reserve in 1976 had a probable increase of 35 % inside the protected area and a decrease of 13 % outside. Thus use of the model suggests that both the Ruaha and the Selous were undergoing compression of elephants in that period. Since that period, major declines have occurred in all these protected areas (Borner and Severre, 1984; Douglas-Hamilton et al., 1986; Dublin and Douglas-Hamilton, 1987)