14/3. Consider now what will happen if a polystable system be subjected to an impulsive (S. 6/5) stimulus S repetitively, the stimulus being unvarying, and with intervals between its applications sufficiently long for the system to come to equilibrium before the next application is made. By S. 6/5, the stimulus S, being impulsive, will displace the representative point from any given state to some definite state. Thus the effect of S (acting on the representative point at a state of equilibrium by the previous paragraph) is to transfer it to some definite state in the field and there to release it. The possibilities sketched in Figure 14/3/1 will illustrate the process sufficiently. Suppose the system is in equilibrium at A. S is applied; its effect is to move the representative point to the end of the arrow, in this example moving it into another confluent. The system is now, by hypothesis, left alone until it has settled: this means that the basic field operates, carrying it, in this example, to the state of equilibrium B. Here it will remain until the next application of S, which in this example, again moves it to a new confluent; here the basic field takes it to the state of equilibrium C. So does the alternation of S and the basic field take it from equilibrium to equilibrium till it arrives at E. From this state, S moves it only to within the same confluent and the ‚leaving alone‘ results in its coming back to E. S (having by hypothesis a unique effect) now takes it to the arrow head, and again it comes back to E. This state of affairs is now terminal, and the representative point is trapped within the E-confluent.
It can now be seen that the process is selective; the representative point ends in a confluent such that the S-displacement carries it to some point within the confluent. Confluents such as A, C, and D, with the S-displacement going outside, cannot hold the representative point under the process considered; confluents such as E, J, and L can trap it.