If the Reykjanes Ridge is supposed to have convection rolls at each side, then they can be extrapolated farther. The mantle is half molten and does not behave like the rigid surface. Therefore, the convection rolls must be continuous. Then, the same shape must prevail all the way.
The Mid Atlantic Ridge itself makes some side-jumps, and is swaying a bit irregularly. Still, the calculated convection rolls division line does cross the ridge itself at 32°N and at equator. This is something to think about, because the system is found to converge around 32°N, whereas the convection rolls there are oriented exactly north-south. At 0°, the fact that horizontal flow Coriolis effect shifts from right hand turn to left hand turn, explains the consistency of distance between main topographic features. In this case, the distance from center of Atlantic is exactly 30° to both S-America and Africa. That means of course, that the width from coast to coast of the Atlantic Ocean along equator is exactly 60°. Not strange, because each convection roll of lower mantle spans 30° from east to west.