a flame five or six inches long (Fig. 53). You might also
have noticed that when the bubble was removed, the vapour inside it
began to pass out again and fell away in a heavy stream, but this you
could only see by looking at the shadow upon the screen.

[Illustration: Fig. 53]

You may have noticed when I made the drops of oil in the mixture of
alcohol and water, that when they were brought together they did not at
once unite; they pressed against one another and pushed each other away
if allowed, just as the water-drops did in the fountain of which I
showed you a photograph. You also may have noticed that the drops of
water in the paraffin mixture bounced against one another, or if filled
with the paraffin, formed bubbles in which often other small drops, both
of water and paraffin, remained floating.

In all these cases there was a thin film of something between the drops
which they were unable to squeeze out, namely, water, paraffin, or air,
as the case might be. Will two soap-bubbles also when knocked together
be unable to squeeze out the air between them? This you can try at home
just as well as I can here, but I will perform the experiment at once. I
have blown a pair of bubbles, and now when I hit them together they
remain distinct and separate (Fig. 54).

[Illustration: Fig. 54.]

I shall next place a bubble on a ring, which it is just too large to get
through. In my hand I hold a ring, on which I have a flat film, made by
placing a bubble upon it and breaking it on one side. If I gently press
the bubble with the flat film, I can push it through the ring to the
other side (Fig. 55), and yet the two have not really touched one
another at all. The bubble can be pushed backwards and forwards in this
way many times.

[Illustration: Fig. 55.]

I have now blown a bubble and hung it below a ring. To this bubble I can
hang another ring of thin wire, which pulls it a little out of shape.
Since the pressure inside is less than that corresponding to a complete
sphere, and since it is greater than that outside, and this we can tell
by looking at the caps, the curve is part of one of those represented by
the dotted lines in C or E, Fig. 31. However, without considering the
curve any more, I shall push the end of the pipe inside, and blow
another bubble there, and let it go. It falls gently until it rests
upon the outer bubble; not at the bottom, because the heavy ring keeps
that part out of reach, but along a circul