Abstract
It is well known that retrieving sequence from long-term memory approximately follows the exponential distribution. The initial divergence from this distribution that occurs frequently supposed to be a result of activation in a single retrieval. Therefore investigating the process of activation is greatly useful to derive an equation which represents the retrieval process with activation. Usual retrieval experiments were carried out to find two parameters changed systematically when data adopted to make up theoretical curves were replaced systematically, and a new equation was derived, from which experimental data showed little divergence. It also predicted the latter half of the sequence from the data of the first half.
