A basic problem when calculating reaction rates using the reactive flux method is the introduction of a reaction coordinate. In this paper we show that it is advantageous to define a reaction coordinate by means of the unstable normal mode of the saddle point of the potential energy surface. This particular choice is made since it yields a high transmission function. Moreover, the reaction coordinate is calculated via a rapidly converging algorithm, and its derivative, which is needed in constrained runs, is calculated analytically. Calculations on the transmission coefficient of the isomerization of n-butane are in good agreement with results published by others. Runs with an isomerizing calixarene in vacuo produce a very high transmission coefficient, as is the purpose of the reaction coordinate. The same molecule is also studied in chloroform.