We start from an assumption that merging occurs in regions of the magnetopause where current strengths are greater than some threshold value which corresponds to the total jump in the field across the magnetopause greater than 50 nT. Because time and cost constraints preclude running numerical simulations for a wide variety of interplanetary magnetic field (IMF) orientations to determine these locations, we adopt an analytical model based on previously derived formulations for magnetospheric and magnetosheath magnetic fields. The magnetospheric magnetic field is confined within a paraboloid. The magnetosheath magnetic field is derived from that in the solar wind and lies between the magnetopause and a paraboloid bow shock. We allow a slight diffusion of the magnetosheath magnetic field into the magnetosphere. The results of the model show that during periods of due southward IMF orientation, merging occurs (as expected) in a wide region centered on the subsolar magnetopause. During periods of northward IMF, connection continues near the subsolar point but also poleward of the cusps. Magnetic energy is only released to the plasma in the latter regions. During periods of strongly northward IMF (By = 0), reconnection ceases on the subsolar magnetopause but continues poleward of the cusp. If the IMF points northward but By is nonzero, reconnection continues near the subsolar point and poleward of the cusps. During periods of sunward IMF orientation, merging nearly ceases on the northern hemisphere (except in the vicinity of the subsolar point) but continues outside the southern lobes. Dawnward and duskward IMF orientations produce tilted patches of enhanced current densities in the subsolar region. We compare the results of our model with previous predictions of the “component” and “antiparallel” merging models.