Since the 1980s, nonlinear dynamic modelling is becoming a popular methodology in economics. However, it is not as new as many researchers seem to believe. Before the linear approach dominated economic theory around the 1950s, many economists were actively involved in the development of nonlinear models, this tendency being particularly strong during the period 1930-1950. The main objective of this essay is to offer a systematic and comprehensive survey of the early developments in nonlinear dynamics in economics, ranging form Frisch’s original impulse and propagation model in 1933, to Goodwin’s formalisation of the limit cycle in 1951.