In our previous experiments we have seen that in order to produce large drops of a given liquid, the surroundings should be of nearly the same density, so as largely to diminish the effective weight of the suspended mass. We might therefore expect that large columns of liquid could be produced under similar conditions; and our conjecture is correct. We may, for example, use the apparatus by means of which large drops of orthotoluidine were formed (Fig. 13), using a shallow layer of water, so that the lower end of the drop would come into contact with the bottom of the vessel before the breaking stage was reached, and thus produce, on a large scale, the same result as that we have just achieved by allowing a hanging drop of water to touch a glass plate. This method, however, restricts the diameter of the top of the column to that of the delivery tube, and in this respect the shape is strained. The remedy for this is to hang the drop from the surface of the water, when a degree of freedom is conferred upon the upper part, which enables the column to assume a greater variety of shapes. In order to show how this may be accomplished, I pour a small quantity of water into the rounded end of a wide test-tube, which is now seen projected on the screen, and then pour gently down the side a quantity of _ethyl benzoate_, a liquid slightly denser than water. You observe that the liquid spreads out on the surface of the water, forming a hanging drop which at first is nearly hemispherical in shape; but as I continue to add the liquid the drop grows in size downwards, and finally reaches the bottom of the tube. There is our liquid column (Fig. 27), which has formed itself in its own way, free from the restriction imposed by a delivery tube. Notice the graceful curved outline, produced by a beautiful balance between the forces of surface tension and gravitation; and notice also how the outline may be varied by the gradual addition of water, which causes the upper surface to rise, and thus stretches the column (Fig. 28). The middle becomes more and more narrow (Fig. 29), and finally breaks across, leaving a portion of the former column hanging from the surface, and the remainder, in rounded form (Fig. 30), at the bottom of the tube. And, as usual, the partition was accompanied by the formation of a small droplet. [Illustration: __Figs._ 27, 28, 29, 30.—A liquid column stretched upwards until broken by addition of water. Four stages._]