Upon application of the theory of stochastic processes to reactor kinetics, it is possible to show that the square of the modulus of the reactor transfer function is proportional to the Fourier transform of the autocorrelation function for power noise in the reactor. Since the power noise represents the response to the minimum power input signal, measurements of transfer functions based upon reactor noise are of all possible measurements, least subject to nonlinear distortion. By performing the experiment at various power levels and temperatures, it is possible to measure both power and temperature coefficients. If the reactor is periodically monitored during its operation, long term changes can also be measured.