PEIRCE. — MAGNETIZATION IN IRON. 127 



with a standard amperemeter in each branch. For the prehminary 

 experiments, fields stronger than 4600 gausses were not needed. 



The thick, sohd-drawn brass tube upon which the wire was wound 

 carried a stream of tap water to keep the specimen at a constant 

 temperature. The test coil was wound upon the test piece after the 

 latter had received a very thin film of varnish. The test coil, after 

 it had been made, was varnished and the whole was then placed for 

 about half an hour in a stream of hot air to harden the coating. The 

 leads were enclosed in a very thin tube of rubber, the test coil was 

 covered with a rubber shield, and melted paraffine wax was then run 

 into the ends of this shield so as to keep the test coil absolutely dry. 

 In this manner all leakage from turn to turn of the triply silk covered 

 wire of which the test coil was made was avoided. In many cases 

 two test coils were wound side by side upon each specimen, but the 

 results, after we had learned how to make the coils properly, were so 

 nearly identical for both coils that we sometimes used only one. In 

 all cases the differences, if there ever were any real differences, were 

 far smaller than the unavoiflable errors of ballistic galvanometer 

 reading. 



The reversal of a strong current in the circuit of a solenoid with 

 so great an inductance as this one, has to be managed carefully. 

 The main reversing switch, when it was slightly pulled, automatically 

 put the solenoid in parallel with a noninckictively wound resistance 

 higher than its own, and, after the handle was raised higher, broke 

 the main circuit so that the discharge from the solenoid could pass 

 through the auxiliary resistance. The process was inverted when the 

 switch handle was pushed down. This switch (Figure 3, Plate 1) was 

 designed and made by Mr. Coulson. 



To prove that the field in the solenoid, when a given current passes 

 through the circuit, is really- what it should he, according to the cal- 

 culation, a very carefully made test coil without iron was placed in 

 series wath the secondary of a standard of mutual inductance and the 

 field was thus measured. By this means it was shown that there was 

 no appreciable leakage between the turns of the solenoid — a very 

 common fault of the exciting coils of electromagnets — and that there 

 was not enough iron in the brass of the reel to affect the field strength 

 sensibly. So far as we can determine the fact by our many and re- 

 peated tests, the solenoid has not been injured by use and is very 

 perfect. It is firmly mounted on a solid oak frame so that its axis is 

 horizontal and perpendicular to the meridian. 



The dimensions of the iron test pieces and of the standard indue- 



