And its levity may be shown very prettily by a simple experiment. Select two gas jars of the same size, and after filling one with oxygen gas and the other with nitrogen gas, slide glass plates over the bottoms of the jars, and proceed to invert the one containing oxygen, placing the neck in a stand formed of a box open at the top; then place the jar containing nitrogen over the mouth of the first, withdrawing the glass plates carefully; and if the table is steady the top gas jar will stand nicely on the lower one. Then (having previously lighted a taper so as to have a long snuff) remove the stopper from the nitrogen jar and insert the lighted taper, which is immediately extinguished, and as quickly relighted by pushing it down to the lower jar containing the oxygen. This experiment may be repeated several times, and is a good illustration of the relative specific gravities of the two gases, and of the importance of the law of universal diffusion already explained at p. 6, by which these gases _mix_, not _combine_ together, and the atmosphere remains in one uniform state of composition in spite of the changes going on at the surface of the earth. Omitting the aqueous vapour, or steam, ever present in variable quantities in the atmosphere, ten thousand volumes of dry air contain, according to Graham:-- [IIlustration: Fig. 108. A. Gas jar containing nitrogen, N, standing on B, another jar full of oxygen, O. The taper, C, is extinguished at N, and relighted at O. D D. Stand supporting the jars.] [Page 106] Nitrogen 7912 Oxygen 2080 Carbonic acid 4 Carburetted hydrogen (CH_{2}) 4 Ammonia a trace ------- 10,000 _Fourth Experiment._ It was the elegant, the accomplished, but ill-fated Lavoisier who discovered, by experimenting with quicksilver and air, the compound nature of the atmosphere; and it was the same chemist who gave the name of azote to nitrogen; it should, however, be borne in mind that it does not necessarily follow because a gas extinguishes flame that it is a _poison_. Nitrogen extinguishes flame, but we inhale enormous quantities of air without any ill effects from the nitrogen or azote that it contains; on the other hand, many gases that extinguish flame are _specific poisons_, such as carbonic acid, carbonic oxide, cyanogen, &c. Lavoisier's experiment may be repeated by passing into a measured jar, graduated into five equal volumes, four measures of nitrogen and one measure of oxygen; a glass plate should then be slid over the mouth of the vessel, and it may be turned up and down gently for some little time to mix the two gases, and when the mixture is tested with a lighted taper, it is found neither to increase nor diminish the illuminating power and the taper burns as it would do in atmospheric air. (Fig. 109.)