would press upon the plate; but the plate would be able to resist the pressure, and the bubble would remain a hemisphere with a flat base. If, on the other hand, the bubble were formed on the surface of a liquid, there would be precisely the same pressure on the bottom, only it would be acting on a medium which would give way to it; the liquid, therefore, would yield to the pressure of the air, and we should have the bubble as it were a little buried in the liquid by its own pressure. As the pressure increases with the smallness of the bubble, we should expect a small bubble to be very deeply buried, and a large bubble to be slightly buried. I will now pour into the cell, the image of which you see, a small quantity of liquid, and blow in it a very small bubble. You now see the images of two bubbles which have risen to the surface, and that they are very much buried in the liquid by virtue of their pressure. I will now blow a large bubble. You see that within it the surface of the liquid is very much less depressed. I will blow a still larger one. Now I have succeeded in blowing a very large bubble, and the lower part of it is not appreciably depressed. I will now blow a great number of bubbles in contact, and will then point out one or two facts. You now see that odd network which represents a great number of bubbles. There are two points I wish you to notice. In the first place, when two bubbles meet, the surface between them may be either plane or curved. It is plane if both bubbles are of equal size, and therefore compress the air within them with equal force; but, if they are unequal, the smaller bubble, compressing the air more strongly, indents the larger, and the surface which divides them is curved. Notice also another very curious point, namely, that in no case do more than three bubbles meet in a point, excepting for an instant. This follows from the law that a large number of bubbles, as well as each one, will assume the smallest possible surface. I cannot go into the proof of this, but it follows from the law I have already given you. As the bubbles form, collapse, and disappear, you see that they always so arrange themselves that no more than three shall ever meet in a point.
Now, then, we have got our bubble on the surface of the liquid. Let us consider what will happen to it after that. Evidently the liquid of which it is composed will run down the sides by virtue of its own weight; but there will be a certain resistance to this motion, greater or less as the viscosity of the surface is great or small. Hence, there are two different dangers which may beset the bubble. The first of these is, that when the surface-viscosity is small, then the liquid runs down the sides of the bubble very easily; the consequence is, the bubble becomes very thin and bursts. There is, however, an opposite danger which may imperil the bubble when the surface-viscosity is great; and that is, that the liquid does not flow down in a straight line or regular curve, but in irregular masses, which every now and then tear away from each other. Now, these ruptures make
