304 The Rev. T. R. Robinson's Experimental Researches on the 



thought sufficient. But it sometimes happens that though each set is perfectly 

 consistent, the two differ as much as 20 lbs., which occurs especially when the 

 magnet's lift is about half its maximum, at which point the coercive force of the 

 iron seems to make some abrupt change. In these cases other sets were taken, 

 till from the uniform spread of the differences I felt satisfied that I had obtained 

 a fair average. 



The third column, headed F^ contains the values of the currents expressed 

 in the unit which has been described above. It must be remembered, however, 

 that they act on 638 spu-es. I consider them true to O'OOl at least of their as- 

 signed amount. The negative sign indicates that in these instances the direction 

 of the current is reversed in the helices. 



The fourth column gives T, the temperature of the magnet, as shown by a 

 thermometer dipped in mercury, which filled the upper inch of the cavity in 

 the northern cylinder of the magnet ; at first both cylinders were tried, but this 

 was foimd useless. It is necessary to know the temperature, for the force of 

 electro-magnets, as of common ones, varies with it. To investigate the correc- 

 tion, 40 feet of leaden pipe, §-incli external, and ^ internal diameter, were 

 coiled on helices, containing 316 spires of the same wire used in the others, but 

 coiled on tin tubes. These worms had each 25 turns ; they were covered 

 with thick cloth, and connected by a tube of vulcanized caoutchouc with a 

 small boiler, so that a current of steam could be passed through them, and the 

 condensed water escaped from their open extremity.* As the keeper and base 

 (which were also covered with cloth) presented much cooling surface, the 

 temperature could not be maintained above 180°, but could be kept very steady. 

 The lift of the magnet being then determined, the magnet was left to cool, 

 and the observation was repeated at the ordinary temperature, and with the 

 same current as nearly as could be managed. As I was not aware of any rea- 

 son for supposing that the effect of temperature changes its law under that of 

 boiling water, I assumed the change to be as the temperature, or, L being the 

 lift at 60°, 



l^L^\l H-T(r-60)|. 



* To the last it was turbid with sulphuret of lead, so that this material cannot be depended on 

 as a conductor of steam. 



