gradually recovers its spherical shape during its fall through the water. Finally, I increase the discharge to such an extent that the formation of the terminal drops is so rapid as to be no longer visible to the eye, but appears like the turmoil observed at the end of a jet of water escaping into air. [Illustration: __Figs._ 23, 24, 25.—Jets of Orthotoluidine, discharged under water._] [Illustration: __Fig._ 26.—Water stretched between a tube and a plate._] *Liquid Columns.*—A simple experiment will suffice to illustrate what is meant by a liquid column. Here is a drop of water hanging from the end of a glass tube. I place it in the lantern and obtain a magnified image on the screen, and then bring up a flat plate of glass until it just touches the suspended drop. As soon as contact is established, the water spreads outwards over the plate, causing the drop to contract in diameter at or near its middle part, so that its outline resembles that of a capstan (Fig. 26). The end of the glass tube is now connected to the plate by a column of water of curved outline, which is quite stable if undisturbed. If, however, I gradually raise the tube, or lower the plate, the narrow part of the column becomes still narrower, and finally breaks across. In the same way we may produce columns of other liquids; those obtained with viscous liquids such as glycerine being capable of stretching to a greater extent than water, but showing the same general characteristics. A liquid column, then, is in reality a supported drop, and the severance effected by lowering the support is similar to that which occurs when a pendent drop breaks away owing to its weight.