ROBIN ROOSTS. Loe 
total number twelve hundred, or thereabout, 
on the assumption that my outlook had cov- 
ered a quarter of the circuit. After the 
flight ceased I went into the wood, and from 
the commotion overhead it was impossible not 
to believe that such a calculation must be 
well within the truth. 
The next day was rainy, but on the even- 
ing of the 28th I stood by the shore of the 
pond, on the eastern side of the wood, and 
made as accurate a count as possible of the 
arrivals at that pomt. Unfortunately I was 
too late; the robins were already coming. 
But in fifty minutes, between 6.40 and 7.30, 
I counted 1072 birds. They appeared singly 
and in small flocks, and it was out of the 
question for me to make sure of them all; 
while I was busy with a flock on the right, 
there was no telling how many might be 
passing in on the left. If my observations 
comprehended a quarter of the circle, and 
if the influx was equally great on the other 
sides (an assumption afterward disproved), 
then it was safe to set the whole number 
of birds at five thousand or more. Of the 
1072 actually seen, 797 came before the sun- 
set gun was fired, —a proportion somewhat 
