Dictyostelium amoebae use chemoattractant-mediated motion to aggregate as a response to starvation. Recent experiments have focused attention on the signaling mechanisms underlying this behavior and on the threshold for gradient detection. This talk reviews the current experimental situation and then compares and contrasts several conceptual models put forth to understand how the system works. A particular interesting question concerns the role of receptor noise in limiting the accuracy of the response, and some preliminary results regarding this issue will be presented.