| Abstract: | ABSTRACT. Statistical issues relating to a number of sampling methods which can be used to obtain estimates of the number, or density, of forest elephants are discussed. The most widely used method, distance sampling, is briefly described. A recently proposed method, recce sampling, that reduces sampling effort is described and discussed. Adaptive sampling, often used when populations are sparse, patchy or clustered, is described and proposed as an alternative. Model‐based and design‐based inferences are compared and their use in sampling forest elephant populations is discussed. Further considerations that are required to estimate or monitor changes in elephant populations over time are outlined. |
