The average patterns of the electrostatic potential, current vectors, and Joule heating in the polar ionosphere, as well as the associated field‐aligned currents, are determined for a quiet time, the growth phase, the expansion phase, the peak epoch, and the recovery phase of substorms. For this purpose, the Kamide‐Richmond‐Matsushita magnetogram‐inversion algorithm is applied to a data set (for March 17, 18, and 19, 1978) from the six meridian magnetometer chains (the total number of magnetometer stations being 71) which were operated during the period of the International Magnetospheric Study (IMS). This is the first attempt at obtaining, on the basis of individual substorms, the average pattern of substorm quantities in the polar ionosphere for the different epochs. The main results are as follows: (1) The substorm‐time current patterns over the entire polar region consist of two components. The first one is related to the two‐cell convection pattern, and the second one is the westward electrojet in the dark sector which is related to the wedge current. (2) Time variations of the two components for the four substorm epochs are shown to be considerably different. (3) The dependence of these differences on the ionospheric electric field and the conductivities (Hall and Pedersen) is identified. (4) It is shown that the large‐scale two‐cell pattern in the electric potential is dominant during the growth phase of substorms. (5) The expansion phase is characterized by the appearance of a strong westward electrojet, which is added to the two‐cell pattern. (6) The large‐scale potential pattern becomes complicated during the recovery phase of substorms, but the two‐cell pattern appears to be relatively dominant again during their late recovery as the wedge current subsides. These and many other earlier results are consistent with the present ones, which are more quantitatively and comprehensively demonstrated in our global study. Thus the two components are tentatively identified as the directly driven and the unloading components, respectively, although there is some indication that both components are actually coupled in the ionosphere. In the present paper we show that the directly driven component is present throughout the lifetime of substorms, becoming a dominant feature during the recovery phase of substorms as the unloading component wanes. The fact that the two components exist and that their time variations are different indicates that different processes are in progress even for the same value, say, −500 nT, ofALand that we must be cautious in using a single electrojet index, such asAL, in