The electrodynamics of a substorm are analyzed using a crude one-dimensional model. The discussion starts from the ''standard interpretation'' scenario in which the plasma sheet collapses into a neutral sheet and magnetic merging occurs between the two tail lobes. Plasma then flows into the neutral sheet from both the lobes and the sides and is accelerated in the dawn-dusk direction according to a ''type 2 merging'' scenario. The lobes supply the process with magnetic energy, but most particles come from the original plasma sheet population, as evidenced by the relatively low acceleration: Alfv¿n's theory for this is modified to accommodate both particle sources. The process is further modified by the tendency of accelerated plasma to unbalance charge neutrality. To maintain such neutrality, it is assumed that the ionosphere absorbs any excess of electrons and provides electrons to neutralize any excess of ions.

It is this exchange of electrons with the ionosphere, rather than some unspecified ''current disruption,'' that accounts for the ''current wedge'' observed to accompany substorms. The cross-tail current is weakened by the diversion, and this reduces the adjacent lobe field intensity, but there are no drastic consequences, either in the wedge circuit or in the somewhat similar diversion of region 1 Birkeland currents: at most the tail boundary may expand very slightly. The accelerating electric field of the merging region is inductive and could in principle be completely confined to the tail. The ionosphere then comes into play only because the wedge currents flow through it and produce a resistive voltage drop, a true electric potential which unlike the inductive field propagates along field lines and ultimately affects the acceleration region. The system is self-stabilizing to some extent, because if the ionospheric potential blocks off plasma sheet inflow into the merging region, it also cuts off the wedge current which produce this potential.

However, Alfv¿nic delays modify this feedback, and a blocking potential might well explain bursts of high-energy particles that are occasionally observed.