Now take another step. Oxygen, as I showed above, three years ago in the _Philosophical Magazine_ for 1847, vol. xxxi., pp. 410, 415, 416, is magnetic in relation to nitrogen and other gases. E. Becquerel, without knowing of my results, has confirmed and extended them in his paper of last year, and given certain excellent measures. In my paper of 1847 I showed also that oxygen (like iron and nickel) lost its magnetic power and its ability of being attracted by the magnet when heated (p. 417). And I further showed that the temperatures at which this took place were within the range of common temperature, for the oxygen of the air--_i.e._ the air altogether--is increased in magnetic power when cooled to 0° F. (p. 406). Now I must refer you to the papers themselves for the (to me) strange results of the incompressibility (magnetically speaking) of oxygen and the inexpansibility of nitrogen and other gases; for the description of a differential balance by which I can compare gas with gas, or the same gas at different degrees of rarefaction; for the determination of the true zero, or point between magnetic and diamagnetic bodies; and for certain views of magnetic conduction and polarity. You will there find described certain very delicate experiments upon diamagnetic and very weak magnetic bodies concerning their action on each other in a magnetic field of equal force. The magnetic bodies repel each other, and the diamagnetic bodies repel each other; but a magnetic and a diamagnetic body _attract_ each other. And these results, combined with the qualities of oxygen as just described, convince me that it is able to deflect the lines of magnetic force passing through it just as iron or nickel is, but to an infinitely smaller amount, and that its power of deflecting the lines varies with its temperature and degree of rarefaction. [Sidenote: ATMOSPHERIC MAGNETISM.]