20 THE ARCHITECTURE OF BIRDS. 
circumference with its claws, and works with its 
bill from the centre outwards; and hence it is that 
in the numerous excavations recently commenced, 
which we have examined, we have uniformly found 
the termination funnel-shaped, the centre being al- 
ways much more scooped out than the circumfer- 
ence. The bird consequently assumes all positions 
while at work in the interior, hanging from the roof 
of the gallery with its back downward as often as 
standing on the floor. We have more than once, 
indeed, seen a bank-swallow wheeling slowly round 
in this manner on the face of a sandbank when it 
was just breaking ground to begin its gallery.* 
This manner of working, however, from the cir- 
cumference to the centre, unavoidably leads to irreg- 
ularities in the direction, which would not so read- 
ily occur by reversing the procedure; for though 
the radius formed by a part of the bird’s body is 
subject to little variation, yet the little that does oe- 
cur from the extension or contraction of the neck, 
must tend to throw it out of the right line. Accord- 
ingly, all the galleries are found to be more or less 
tortuous to their termination, which is at the depth 
of from two to three feet, where a bed of loose hay 
and a few of the smaller breast-feathers of geese, 
ducks, or fowls are spread with little art for the re- 
ception of the eggs.f 
It may not be unimportant to remark also, that it 
always scrapes out with its feet the sand detached 
by the bill; but so carefully is this performed, that it 
never scratches up the unmined sand or disturbs the 
plane of the floor, which rather slopes upward, and, 
of course, the lodgment of rain is thereby prevented. 
The bank-swallow is eminently a social bird; since 
it not only always nestles in numerous colonies, but 
also hunts for insects in troops of from three to fifty, 
and, as Buffon correctly remarks, assocjates freely 
with other swallows. 
* J. Rennie. + J. Rennie. 
