To see why, assume that the crystallographic orientation (i.e., where the hexagonal surface sits in relation to the car hood surface) is completely determined by something called the “c-axis". (Strictly speaking, two axes are needed, but the second one here is unimportant.)

In the center, the c-axis points straight up, off the hood. Charles Knight found in his experiments that the initial ice growth in the film grows outward along its branches such that the c-axis of newer ice regions are more nearly straight up. The curvature of the branches also relates to the tilt of the c-axis. On each eyelid, the axis tilts toward the center region. In the center region the tilt lies in-between, which is straight up, likes poles leaning against a flagpole.

This doesn’t explain the pattern. It just tells us why the hexagonal forms lie in the middle of the eye.

Too bad there wasn’t a circle cut in half under the eye. When I find that one, we’ll have “frost scream". Actually, that’s a bit hard to say.

- Mark