the road a smoothly curved expanse covered the whole of
the road allowance and gently sloped down into the field at my left. Its
north edge stood like a cliff, the exact height of the fence, four feet
I should say. In the centre it rose to probably six feet and then fell
very gradually, whaleback fashion, to the south. Not one of the fence
posts to the left was visible. The slow emergence of the tops of these
fence posts became during the following week, when I drove out here
daily, a measure for me of the settling down of the drift. I believe I
can say from my observations that if no new snow falls or drifts in,
and if no very considerable evaporation takes place, a newly piled
snowdrift, undisturbed except by wind-pressure, will finally settle down
to about from one third to one half of its original height, according
to the pressure of the wind that was behind the snow when it first was
thrown down. After it has, in this contracting process, reached two
thirds of its first height, it can usually be relied upon to carry horse
and man.

The surface of this drift, which covered a ditch besides the grade and
its grassy flanks, showed that curious appearance that we also find in
the glaciated surfaces of granite rock and which, in them, geologists
call exfoliation. In the case of rock it is the consequence of extreme
changes in temperature. The surface sheet in expanding under sudden heat
detaches itself in large, leaflike layers. In front of my wife's cottage
up north there lay an exfoliated rock in which I watched the process for
a number of years. In snow, of course, the origin of this appearance
is entirely different; snow is laid down in layers by the waves in the
wind. "Adfoliation" would be a more nearly correct appellation of the
process. But from the analogy of the appearance I shall retain the more
common word and call it exfoliation. Layers upon layers of paperlike
sheets are superimposed upon each other, their edges often "cropping
out" on sloping surfaces; and since these edges, according to the
curvatures of the surfaces, run in wavy lines, the total aspect is very
often that of "moire" silk.

I knew the road as well as I had ever known a road. In summer there was
a grassy expanse some thirty feet wide to the north; then followed the
grade, flanked to the south by a ditch; and the tangle of weeds and
small brush beyond reached right up to the other fence. I had to stay
on or rather above the grade; so I stoo