comparatively low temperatures and with small magnetising forces, yet, from the contradictory results which had been obtained by other experimenters, directly opposite conclusions as to the magnetic intensities of the land and ocean areas respectively might, with some show of reason, be drawn from those which I had arrived at. The important bearing which the influence of temperature has upon the phenomena of terrestrial magnetism, induced me to undertake an investigation into the causes of the conflicting results hitherto obtained, with the hope, also, that I might be able to extend still further our knowledge of magnetic substances.

The results of my experiments, which are embodied in a paper read before the Royal Society,* confirm the general law of the diminution of the magnetisation of magnetic substances with increase of temperature for small as well as for large magnetising forces. I have also demonstrated in this paper that the apparent increase of the magnetic power of heated iron, magnetite, and nickel is so small as to be a negligible quantity in general magnetic phenomena and terrestrial magnetism, and is due to a surface resistance of these substances which disappears(1) on the application of heat; (2) by the action of strong magnetising forces; (3) by diminishing the mass of the magnetic substance acted upon by the magnetising force.

I have further shown that the surface resistance of cobalt at normal temperatures is so great, as to require a tractive force equal to 373 lbs. per square inch, acting on a minute quantity of the metal, to overcome it.

The general results of my experiments have been confirmed by M. P. Curie, in two able papers in the Comptes Rendus of the French Academy for April and May of the present year. M. Curie also agrees with my conclusion

Proc. Roy. Soc., June 11, 1891.

that the apparent increase of the magnetisation of iron, magnetite, and nickel is anomalous, and masks the principal phenomena of the decrease of magnetic power with ascending temperatures. M. Curie has extended his observations to the magnetic behaviour of gaseous oxygen, and has found with Professor Dewar, when experimenting with this element in a liquid state, a decrease of power with increase of temperature.

The general law of the diminution of the magnetisation of all known magnetic substances with increasing temperatures is now completely established.

In connexion with this brief summary of experiments on the influence of temperature on magnetic substances, I would direct attention to the close analogy, if not an actual relation, which subsists between the anomalous surface resistance of cold iron to magnetisation and its anomalous property of resisting chemical action. Schönbein and Faraday have shown that bright iron wire, slightly oxidised by heat, is quite insensible to the action of strong nitric acid.* Not only is there no reaction under these conditions, but the oxidised wire has the property of inducing voltaic passivity in a number of pieces of bright iron wire by simple contact with them when immersed in the acid. Further, a passive bright wire has the property of inducing the passive condition in other pieces of ordinary bright iron wire.

It is admitted on all hands that this anomalous voltaic condition of iron is a surface resistance, as it disappears (1) by abrasion; (2) by the action of dilute nitric acid; and (3) by the application of heat to the wire.

To affirm, therefore, as a general property of iron, that its magnetic power increases with the temperature, is as irrational as to maintain that iron throughout its substance is, like gold and platinum, insensible to the action of strong nitric acid.

* Phil. Mag., 1836, Vol. IX., pp. 53-65.

6 Magnetometer for showing the Influence of Temperature.

For demonstrating the influence of temperature on magnetic substances, I have devised a magnetometer which is shown in the accompanying figure (Plate I.) two-thirds the actual size.

The instrument consists of a declination needle freely suspended from a double fibre of untwisted silk over a disk of brass. One end of the needle is thickly covered with spun silk to prevent the weakening of its magnetism by close proximity to the heated substance under examination. The excursions of the needle are limited in both directions by pins inserted a little distance apart on a diamcter near the edge of the disk. The disk is pivotted on the end of an arm, to which it can be clamped firmly by means of a milled screw when the needle is drawn out of the magnetic meridian. Three binding screws are mounted at equal distances from each other round the circumference of a circular table, which has an independent movement round the disk. The magnetic substances are held in loops of platinum wire fixed by the binding screws to the table, and the properties of the specimens can be examined in succession.

The action of the instrument is as follows:-The needle is drawn out of the magnetic meridian from 15° to 20° by turning the disk on its axis. The magnetic substance is then brought round towards the needle until equilibrium is established between its magnetism and the horizontal component of the earth's magnetism. The magnetic substance is heated by a small gas flame from below; when the needle recedes from heated iron, magnetite, and nickel, and advances again when the source of heat is removed; thereby indicating a decrease of magnetic power for these substances. On the other hand, the needle advances towards cobalt when heated, and recedes when the metal is cooled, by reason of its enormous surface resistance, which only disappears, as I have said, under the action of powerful magnetising forces.