"In the summer of 1948 [Walter]  was teaching a servomechanism class at North American Aviation in which all possible ways of solving a quadratic system had been presented. One of the students then asked how these results changed when the system became “slightly cubic.” “Slightly cubic” refers to a servo in which the amplifier delay in developing voltage for the motor is small but not negligible. (N)either the Nyquist diagram [n]or the Bode plot would show the effect of the amplifier time constant on the frequency response of the system. One of the quadratic descriptions, however, had been a plot of the roots in the s-plane. The extra time constant was added to this plot and the roots shifted. This description was instantly preferred by most students in the class. This was the beginning of the root-locus method for finding the roots of servo equation."

 From Into Stability, Part II, Chapter 6, p. 51