We have already seen that a drop of liquid possesses an elastic surface, and is practically the same thing as a soap-bubble filled with liquid instead of air. We might therefore expect the same results if two suspended drops of liquid were placed in communication as those observed in the case of soap-bubbles. And our reasoning is correct, as we may now demonstrate. The apparatus consists (Fig. 32) of two parallel tubes, each provided with a tap, and communicating with a cross-branch at the top, which contains a reservoir to hold the liquid used. About half-way down the parallel tubes a cross-piece, provided with a tap, is placed. We commence by filling the whole of the system with the liquid under trial, and the parallel tubes equal in length. Drops are then formed at the ends of each vertical tube by opening the taps on these in turn, and closing after suitable drops have been formed. Then, by opening the tap on the horizontal cross-piece, we place the drops in communication and watch the result. I have chosen orthotoluidine as the liquid, and by placing the ends of the vertical tubes under water—which at the temperature of the room is slightly less dense than orthotoluidine—I am able to form much larger drops than would be possible in air. You now see a small and a large drop projected on the screen; and I now open the cross-tap, so that they may communicate. Notice how the little drop shrinks until it forms merely a slightly-curved prominence at the end of its tube. It attains a position of rest when the curvature of this prominence is equal to that of the now enlarged drop which has swallowed up the contents of the smaller one. So far the result is identical with that obtained with soap-bubbles; but we can extend the experiment in such a way as to reverse the process, and make the little drop absorb the big one. In order to do this I fasten an extension to one of the tubes, and form a small drop deep down in the water, and a larger one on the unextended branch near the top. When I open the communicating top, the system becomes a kind of siphon, the orthotoluidine tending to flow out of the end of the longer tube. The tendency of the large drop to siphon over is opposed by the superior pressure exerted by the skin of the smaller drop; but the former now prevails, and the big drop gradually shrinks and the little one is observed to grow larger. It is possible by regulating the depth at which the smaller drop is placed, to balance the two tendencies, so that the superior pressure due to the lesser drop is equalled by the extra downward pressure due to the greater length of the column of which it forms the terminus. Both pressures are numerically very small, but are still of sufficient magnitude to cause a flow of liquid in one or other direction when not exactly in equilibrium. In the case of communicating soap-bubbles, containing air and surrounded by air, locating the small bubble at a lower level would not reverse the