the ice was gone, and it was too late to
study the bottom. Being curious to know what position my great bubbles
occupied with regard to the new ice, I broke out a cake containing a
middling sized one, and turned it bottom upward. The new ice had formed
around and under the bubble, so that it was included between the two
ices. It was wholly in the lower ice, but close against the upper, and
was flattish, or perhaps slightly lenticular, with a rounded edge, a
quarter of an inch deep by four inches in diameter; and I was surprised
to find that directly under the bubble the ice was melted with great
regularity in the form of a saucer reversed, to the height of five
eighths of an inch in the middle, leaving a thin partition there between
the water and the bubble, hardly an eighth of an inch thick; and in many
places the small bubbles in this partition had burst out downward, and
probably there was no ice at all under the largest bubbles, which were a
foot in diameter. I inferred that the infinite number of minute bubbles
which I had first seen against the under surface of the ice were now
frozen in likewise, and that each, in its degree, had operated like
a burning-glass on the ice beneath to melt and rot it. These are the
little air-guns which contribute to make the ice crack and whoop.

At length the winter set in good earnest, just as I had finished
plastering, and the wind began to howl around the house as if it had
not had permission to do so till then. Night after night the geese came
lumbering in the dark with a clangor and a whistling of wings, even
after the ground was covered with snow, some to alight in Walden, and
some flying low over the woods toward Fair Haven, bound for Mexico.
Several times, when returning from the village at ten or eleven o'clock
at night, I heard the tread of a flock of geese, or else ducks, on the
dry leaves in the woods by a pond-hole behind my dwelling, where they
had come up to feed, and the faint honk or quack of their leader as they
hurried off. In 1845 Walden froze entirely over for the first time on
the night of the 22d of December, Flint's and other shallower ponds and
the river having been frozen ten days or more; in '46, the 16th; in '49,
about the 31st; and in '50, about the 27th of December; in '52, the 5th
of January; in '53, the 31st of December. The snow had already covered
the ground since the 25th of November, and surrounded me suddenly
with the scenery of winter. I withdr