EFFECT OF TEMPERATURE. 69 



1 of temperature, without depending on the assump- 

 tion of uniformity of the Earth's magnet-power, the 

 disturbing needle is inclosed in a copper box; and 

 water, at different temperatures for the different 

 observations, is poured into the box. At each of 

 these temperatures, the deflexion produced on B is 

 observed, and the thermometer is read. Thus we 



A 

 obtain the values of -~ at different temperatures of A ; 



and, as the experiment can be completed in so short a 

 time that we may presume on the invariability of E 

 during the observations, we find in fact the propor- 

 tionate change of A for a given change of temperature. 



In some instances, the temperature has been 

 changed by heating the air of the room. 



The ratio of change thus found is very different for 

 different magnets. It probably depends on the quality 

 of the steel; or possibly on the mode of magnetization: 

 but on this point nothing is known with certainty. In 

 all instances, it is believed, the magnet-power dimi- 

 nishes when the temperature is raised. In the 

 " Horizontal Force Magnet " of the Royal Observatory, 

 Greenwich, the loss of power for a rise of 1 Fahrenheit 

 is = magnet-power x 0-0000809: in the magnet lately 

 used as " Vertical Force Magnet," the loss of power 

 for 1 is = magnet-power x 0'00013845. 



In using a needle in the manner described in 

 Article 25, - for comparing the terrestrial horizontal 

 force at different localities, the result ought always to 



