thus the air is lighter and more moveable in the former pit, so that the
recoil will make itself more felt in A than in B: accordingly, we found
that the main currents alternated between A and C, with very slight
disturbance in the neighbourhood of B. B will, however, play its part,
and the weighty column of air contained in it will oscillate, though
with smaller oscillations than in the case of A. Probably, when the sun
has left A, while acting still upon C, the return current from C will be
much slighter, and there will be a general settling of the atmosphere in
the pits A and B, until C also is freed from the sun's action, when the
whole system will gradually pass into a state of equilibrium.

With respect to the action of the more protected pits, the principle of
the hydraulic ram not unnaturally suggests itself.

In considering the minor details of the currents, such elements as the
refrigeration of the air in its passage across the face of the ice must
be taken into account. It may be observed that the candle did not occupy
an _intermediate_ position with respect to two opposing currents, for it
was practically on the floor of the cave, owing to the continuity of the
slope of snow on which it stood, as shown in the vertical section on p.
108.]


* * * * *




CHAPTER VIII.