278 Wisconsin Academy of Sciences , Arts and Letters . 
means of this rod they could be moved simultaneously over correspond¬ 
ing portions of the two magnets. At G is a gauge which regulates the 
distance the coils are moved at a time, so that as they are moved step by 
step from one end of the bar to the other, the steps will be of equal 
length. By loosening a screw the coils may be moved from the middle 
of the magnet to either end at one step. 
The cross sections of two cylindrical double boxes made of sheet zinc 
are indicated at S and T. At K K are openings in which corks holding 
thermometers were inserted. At L L are openings into the space be¬ 
tween the two parts of the double box. Through these a current of 
steam or cold water was passed to keep the space containing the magnets 
at the requisite temperature. The temperatures employed were 14° and 
99.5° C. The water used to produce the lower temperature was city water 
direct from the faucet. A fairly constant temperature was easily main¬ 
tained. M N, M N, are openings by which the magnets were introduced 
and through which the bars D D and PP passed. They were about 2.5 
cm. in diameter and 20. cm. long, und were stuffed with cotton, the bet¬ 
ter to maintain the temperature of the interior. 
The exploring coils, an ordinary astatic galvanometer of low resistance, 
an earth inductor, arid a resistance box were connected up in series. At 
first the exploring coils were so joined that the currents produced by 
moving them opposed each other. Beginning at the middle of the mag¬ 
nets the coils were moved step by step to one end, the throw of the needle 
being observed for each step. The coils were moved so far in the lastmea- 
surment that practically no lines of induction passed through them as was 
determined by experiment. Similar observations were made for the other 
half of the magnets. This gives the excess of the number of lines of in¬ 
duction passing from a certain section of one magnet into the air over 
that passing out of a corresponding section of the other magnet, that is 
the difference of distribution of magnetism in the two magnets. These 
measurments are taken first (say) when A and B are both at 14°C, and 
and again when A is at 14° C, but B at 99.°5 C. The difference between 
the two sets after they have been reduced to the same scale by the earth 
inductor readings is evidently the change in distribution in B due to 
the change in the temperature. By this method the quantity observed 
is of about the same magnitude as the quantity we desire to obtain. 
To get the distribution of magnetism of the bars, the difference in dis¬ 
tribution was first measured as above at the lower temperature, then 
the connections of the coils were changed so that the induced currents 
were in the same direction and the sum 'of the distributions was meas¬ 
ured in a similar way. In this case extra resistance had to be added 
from the resistance box to keep the readings on the scale. 
