d nothing out of the common course.
They flew low,--not as if they were starting upon any prolonged
flight,--and a goodly number alighted for a little in the field where I
was standing. Shortly before sunrise I went into the wood and found it
deserted. The robin is one of our noisiest birds. Who would have
believed that an assembly of thousands could break up so quietly? Their
behavior in this regard may possibly have been influenced by prudential
considerations. I have said that many of them seemingly took pains to
approach the roost indirectly and under cover. On the westerly side, for
example, they almost invariably followed a line of bushes and trees
which runs toward the roost along the edge of the meadow, even though
they were obliged sharply to alter their course in so doing.

All this time I had been in correspondence with my friend before
referred to, who was studying a similar roost,[15]--in Belmont,--which
proved to be more populous than mine, as was to be expected, perhaps,
the surrounding country being less generally wooded. It was a mile or
more from his house, which was so situated that he could sit upon his
piazza in the evening and watch the birds streaming past. On the 11th of
August he counted here 556 robins, of which 336 passed within five
minutes. On the 28th he counted 1180, of which 456 passed within five
minutes,--ninety-one a minute! On the 2d of September, from a knoll
nearer the roost, he counted 1883 entries.

[15] This roost was discovered by Mr. William Brewster, in August, 1884,
as already mentioned.

This gathering, like the one in Melrose, was greatly depleted by the
middle of September. "Only 109 robins flew over the place to-night," my
correspondent wrote on the 25th, "against 538 September 4th, 838 August
30th, and 1180 August 28th." Two evenings later (September 27th) he went
to the neighborhood of the roost, and counted 251 birds,--instead of
1883 on the 2d. Even so late as October 9th, however, the wood was not
entirely deserted. During the last month or so of its occupancy, the
number of the birds was apparently subject to sudden and wide
fluctuations, and it seemed not unlikely that travelers from the north
were making a temporary use of the well-known resort. It would not be
surprising if the same were found to be true in the spring. In April,
1890, I saw some things which pointed, as I thought, in this direction,
but I was then too closely occupied to follow the matter.

H