An experiment illustrated in Fig. 5, and successfully performed while writing these lines, will settle all doubts, if any remain, upon this puzzling point.
Attach a pith-ball to a short piece of thread, and, knotting the other end, slip the knot in a slit in one end of a quill so as to secure it firmly, and retain the ball about an inch from the other end of the
quill, as shown in Fig. 5. Now blow through the quill Q steadily, and the ball B can be made to hang from or upon the under side of the jet J, being prevented by the string S from blowing away, and the atmospheric pressure A balancing the gravity G of the light ball. As the atmosphere presses nearly fifteen pounds upon a square inch, a ball one-quarter of a pound in weight would be balanced by the full pressure of the air upon a surface only 1/60 of an inch in area. It remains to notice the rapid revolutions of the ball and its uncertain axis of rotation when first suspended in the jet.
Few balls being perfectly round, or of uniform density, will revolve on their own centres under such conditions, but on the shortest axis passing through their centre of gravity (seldom the centre of form).
It is easy to show that all bodies freely suspended tend to revolve on their shortest axis, by tying a string a yard long to a door-key, just under the head, so that it will hang nearly vertical. If the string