During the examination of liquids a means has accidentally been discovered by which a glass tube may be filled with fluorine gas.  A few liquids, one of which is carbon tetrachloride, react only very slowly with fluorine at the ordinary temperature.  By filling a glass tube with such a liquid, and inverting it over a platinum capsule also containing the liquid, it is possible to displace the liquid by fluorine, which, as the walls are wet, does not attack the glass.  Or the glass tube may be filled with the liquid, and then the latter poured out, leaving the walls wet; the tube may then be filled with fluorine gas, which being slightly heavier than air, remains in the tube for some time.  In one experiment, in which a glass test tube had been filled with fluorine over carbon tetrachloride, it was attempted to transfer it to a graduated tube over mercury, but in inclining the test tube for this purpose the mercury suddenly came in contact with the fluorine, and absorbed it so instantaneously and with such a violent detonation that both the test tube and the graduated tube were shattered into fragments.  Indeed, owing to the powerful affinity of mercury for fluorine, it is a most dangerous experiment to transfer a tube containing fluorine gas, filled according to either the first or second method, to the mercury trough; the tube is always shattered if the mercury comes in contact with the gas, and generally with a loud detonation.  Fluorine may, however, be preserved for some time in tubes over mercury, provided a few drops of the non-reacting liquid are kept above the mercury meniscus.

For studying the action of fluorine on gases, a special piece of apparatus, shown in Fig. 3, has been constructed.  It is composed of a tube of platinum, fifteen centimeters long, closed by two plates of clear, transparent, and colorless fluorspar, and carrying three lateral narrower tubes also of platinum.  Two of these tubes face each other in the center of the apparatus, and serve one for the conveyance of the fluorine and the other of the gas to be experimented upon.  The third, which is of somewhat greater diameter than the other two, serves as exit tube for the product or products of the reaction, and may be placed in connection with a trough containing either water or mercury.

The apparatus is first filled with the gas to be experimented upon, then the fluorine is allowed to enter, and an observation of what occurs may be made through the fluorspar windows.  One most important precaution to take in collecting the gaseous products over mercury is not to permit the platinum delivery tube to dip more than two or at most three millimeters under the mercury, as otherwise the levels of the liquid in the two limbs of the electrolysis U-tube become so different, owing to the pressure, that the fluorine from one side mixes with the hydrogen evolved upon the other, and there is a violent explosion.

[Illustration:  FIG. 3.]