424 Mr. F. H. Pitcher on Efects of Temperature and 



tube, and was made sufficiently long to extend at both ends 

 beyond the brass tube. 



Very fine platinum wires (0*003 centim. diameter) were 

 attached at 15 centim. apart to the iron wire specimen. They 

 served as potential leads, and were brought out beyond one 

 end of the glass containing tube, through sealed capillary 

 tubes. 



The glass tube was made tight at both ends by fitting 

 brass cups over and filling with fusible alloy, one end of the 

 tube having been drawn down so that when capped it could 

 be slipped into the solenoid tube. A copper tube was intro- 

 duced through and soldered in the larger brass cap, to serve 

 for exhausting. The vacuum was maintained by a five-fall 

 Sprengel pump, assisted in the early stages by a water pump. 



The remaining apparatus consisted mainly of resistance- 

 boxes, rheostats, special arrangements of mercury-cup con- 

 tacts, switches, storage-batteries, &c. 



Preliminary Tests. 



The specimen was of commercial so-called soft iron wire, 

 and was carefully annealed and polished before mounting. 

 The vacuum-tube containing it was connected to the Sprengel 

 pump and a high vacuum maintained while the wire was 

 being heated by the current. It was observed by the eye 

 that the heating was very uniform, the whole becoming 

 an even red right up to the ends, at a high vacuum. The 

 zero-point or resistance at 0° C. of the specimen was now 

 obtained. It was then placed in the solenoid and the 

 equivalent magnetic length determined. This was found to 

 be a little over 20 centim., and the magnetometer distance 

 was arranged so that slight changes of the length had a 

 minimum effect on the magnetometer readings. A pre- 

 liminary test for the magnetic quality of the iron at ordinary 

 temperatures w T as first made. It was found before further 

 annealing to be fairly hard, having a hysteresis loss for 

 B = 17,000, of 16,000 ergs per cub. centim., and a permeability 

 at that induction of 500. 



After annealing several times in a vacuum, the loss at 

 nearly the same induction had fallen to 6000 ergs, and finally, 

 after successive annealings, arrived at the extraordinarily low 

 value of 557 ergs for B = 3500 at ordinary temperatures. 

 This, in spite of the fact that the specimen was only com- 

 mercial wire, is almost as good as the best specimen of 

 transformer iron tested by Ewing. By this time the wire 

 had settled down to a very steady magnetic state, as shown 



