Changes in Temperature and Distribution of Magnetism. 277 
200° C., using an oil bath to obtain his high temperatures. The changes 
were quite regular between 0° and 180° C., but were very large near 190° C. 
Between 0° and 180° C., he found that the following formula held: 
M — A £ 1 + K— 1 — K~ x — K ( l + x "> J 
where M is the induction in the magnet at a distance x from one of the 
ends; 1, the length of the magnet; A, a quantity depending only on the 
temperature, while K is sensibly constant for a given magnet. 
II. DESCRIPTION OF APPARUTUS AND EXPERIMENT. 
The determination of the reason for and laws governing the increase 
and decrease in magnetism from changes in temperature forms an inter- 
* esting problem, and the investigation was undertaken in the hope of 
throwing new light on the subject. The apparatus employed was sug¬ 
gested by Prof. Rowland, under whose direction the investigation was 
conducted, and will be best understood from the diagram on plate XI. 
A and B are two cylindrical soft steel magnets (Stub’s steel of the 
same temper as when purchased) 30.1 cm. long, and 0.55 cm. in diameter. 
They were magnetized to saturation in a coil, the magnetic circuit being 
completed by an iron casting, which happened to be of suitable size and 
shape. They were then brought to the permanent state by alternate 
heating and cooling. In both ends of each holes were bored and threads 
cut. The depth of these holes in magnet A was 8 mm. at each end. In 
magnet B the hole at the south end was 11 mm. deep, that at the north 
end 7 mm. In the experiment the magnets were placed perpendicular 
to the earth’s field. 
Pieces of brass rod H H H H of the same diameter as the magnets were 
screwed into their ends and acted as guides for the two coils (to be de¬ 
scribed presently), so that after they had been slipped off the magnets, 
they could be slipped back again without trouble. DD is a brass rod 
about 1.5 metres long, holding the two ‘magnets together in the position 
shown in the figure. CC are pieces of non-conducting material to keep 
the magnets from changing temperature by conduction along the rod 
DD. 
E and P are two coils of very fine wire wound on paper tubes which 
just fit the magnets. They consisted of 150 turns (five layers of thirty 
turns each), and were about 7 mm. wide. Their frames were joined by a 
brass rod PP, of such length that when the coil E was at the center of 
the magnet B, the coil F was also at the center of the magnet A. By 
