to put a narrow tube into this
liquid to show that the level is lower in the tube than in the
surrounding vessel, but the same result may be obtained by having a wide
and a narrow tube joined together. Then, as you see upon the screen, the
quicksilver is lower in the narrow than in the wide tube, whereas in a
similar apparatus the reverse is the case with water (Fig. 10).

[Illustration: Fig. 10.]

I want you now to consider what is happening when two flat plates partly
immersed in water are held close together. We have seen that the water
rises between them. Those parts of these two plates, which have air
between them and also air outside them (indicated by the letter _a_ in
Fig. 11), are each of them pressed equally in opposite directions by the
pressure of the air, and so these parts do not tend to approach or to
recede from one another. These parts again which have water on each side
of each of them (as indicated by the letter _c_) are equally pressed in
opposite directions by the pressure of the water, and so these parts do
not tend to approach or to recede from one another. But those parts of
the plates (_b_) which have water between them and air outside would,
you might think, be pushed apart by the water between them with a
greater force than that which could be exerted by the air outside, and
so you might be led to expect that on this account a pair of plates if
free to move would separate at once. But such an idea though very
natural is wrong, and for this reason. The water that is raised between
the plates being above the general level must be under a less pressure,
because, as every one knows, as you go down in water the pressure
increases, and so as you go up the pressure must get less. The water
then that is raised between the plates is under a less pressure than the
air outside, and so on the whole the plates are pushed together. You can
easily see that this is the case. I have two very light hollow glass
beads such as are used to decorate a Christmas tree. These will float in
water if one end is stopped with sealing-wax. These are both wetted by
water, and so the water between them is slightly raised, for they act in
the same way as the two plates, but not so powerfully. However, you will
have no difficulty in seeing that the moment I leave them alone they
rush together with considerable force. Now if you refer to the second
figure in the diagram, which represents two plates which are neither of
them w