SummaryThe paper concerns the circumstances surrounding the collection of ivory from dead elephants, with particular reference to Murchison Falls National Park.The characteristics of the interval between death and complete disintegration of an elephant are described. These, combined with observations of known age skeletons, comprised the criteria used in classifying skeletons found from the air into three relative age classes.Average annual mortality is estimated for the population north of the Nile (MFPN) at 147 animals yielding 1945 kg of ivory, and for that south of the Nile (MFPS) at 474 animals yielding 7497 kg ivory. Park‐found ivory records are analyzed for the 11 y 1959–69. The expected age distribution of deaths is compared with the observed. For MFPN a bias in favour of large (male) tusks is present, explicable by the concentration of ranger search effort in areas of known high male density. For MFPS a bias towards small tusks is thought to be caused by elephants wounded outside the park dying inside it.The National Park recovers an average of 27.6 % of its available ivory per annum, with large annual fluctuations probably correlated with the incidence of wounding outside the park. High losses to poachers are evident.An aerial search for ivory showed a tendency for elephants to die near watercourses. A finding rate of one carcass every 4.3 km of watercourse was obtained. As only 5 % of carcasses still had tusks the aerial searching was prematurely terminated. The results indicated a finding efficiency of 26.4% of the available current year carcasses. Comparative costing suggests that ground searching would be a more efficient method of finding ivory than aerial searching. The high value of the available ivory in Murchison and other areas justifies intensive searching. The low collection rates prevailing in East Africa are largely attributable to the absence of appropriate search effo