Résumé : Dog thyroid slices have been stimulated in vitro by thyrotropin. The kinetics of adenosine 3':5'-monophosphate (cyclic AMP) accumulation due to the activation of adenylate cyclase were measured. Experimental results have been interpreted by means of a numerical simulation based on a theoretical description of the overall process which consists of three steps: (a) the penetration of thyrotropin in the slices, (b) the binding of the hormone to the specific receptor and the consequent activation of adenylate cyclase, (c) the kinetics of cyclic AMP accumulation in each cell due to the interplay between synthesis and degradation. The numerical values of the parameters needed for the simulation were deduced from separate experiments. Diffusion of thyrotropin in the slices has been calculated from the kinetics of efflux of [3H]sucrose, and with or without albumin. Adenylate cyclase activity and activation by thyrotropin in homogenates and purified membranes were measured. Binding experiments of cyclic AMP on crude extract of dog thyroid lead to the conclusion that the maximal capacity of the specific binding site is close to the cyclic AMP content in resting thyroid cells. The purpose of this paper is to verify the validity of the current concepts about the mechanisms taking place in this process by comparing experimental results and theoretical simulation. It results from the study that for this system the time between the activation of adenylate cyclase and half maximal cyclic AMPaccumulation is equal to 3 min 45 s and the pool of cyclic AMP is renewed three times per minute. It has been also stressed that the thickness of the slices as well as inhibitors of phosphodiesterases change the apparent kinetics of cyclic AMP accumulation.