764< REPORT— 1897. 



steady temperature tliroughout the length of the platinum tube, the whole was 

 inclosed iu a vacuum tube. The vacuum was maintained by a five-fall Sprengel 

 pump, which was kept running during the experiment and was assistid by a water- 

 pump in the early stages of exhausting the tube. 



With this method in view, a platinum tube was constructed from a strip of 

 foil 20 cm. long, 2-95 cm. wide, and 0-00254 cm. thick. It was rolled around a 

 mandrel 0'438 cm. diam., and after annealing kept its dimensions without necessi- 

 tating binding wires. The fine platinum leads were attached by winding their 

 ends once around the tube and twisting up tight. A specimen of iron wire was 

 threaded through mica wads of the same diameter as the mandrel, and slipped 

 into position inside the platinum tube. The ends of the platinum tube were bound 

 with bare copper wire to J copper rods, and the whole carefully centred in a 

 glass tube. 



The tube was exhausted, and a pi-eliminary test for uniformity of heating made 

 before placing it in the solenoid. It was seen to be heated very uniformly up to 

 2 cm. from each end. The apparatus was then fixed in position in the solenoid, 

 and the whole placed in position with respect to the magnetometer. It was found, 

 however, that platinum tubes constructed in this manner from thin foil would not 

 stand at high temperatures. Two others were tried, which, on account of not 

 having thicker foil of sufficient length, had to be constructed out of foil of one-half 

 the thickness — one being wound three times around the mandrel and the other five 

 times. These shared the same fate as the first one, giving way in circular cracks 

 running sometimes over one-half the width of the foil. 



The initial exten.sion of the copper springs was not more than 2 mm. in any 

 case, and the ends of the tubes made as square as possible so as to equalise the 

 tension. This method was therefore abandoned until drawn tubes of suitable 

 thickness could be obtained. In the meantime the platinum tube was replaced by 

 the iron wire specimen itself, and in this way observations were obtained, and, by 

 making the specimen its own thermometer, very exact values of its temperature 

 could be obtained. 



By heating the specimen in this manner there would be a strong circular field 

 due to the heating current. The curves plotted from the temperature and magneto- 

 meter readings would therefore have to be corrected for this circular magnetism. 

 This was effected in the manner explained, which, however, limited the temperature 

 to which the wire could be raised. A set of observations was taken with the 

 specimen in air and another in a high vacuum. The true magnetic behaviour of 

 the iron with temperature could then be obtained from these two sets. 



Descrijytion of Affaratus. 



The Solenoid. — The magnetising coil as made in the laboratory was wound on 

 a brass tube about 70 cm. long and having an outside diameter of 2-23 cm. The 

 tube was fairly straight, and previous to winding was filed up and polished in a 

 lathe and then carefully lacquered. 



Windw(/. — ■ 



Length of winding 60-25 cm. 



Depth ,, (4 layers double silk covered It 24 B. and S.) 0-245 „ 



Whole number of turns 40790 „ 



Total resistance 28-1 



This winding was fed on by hand in a lathe and was very fairly uniform. Each 

 layer was carefully paraffined before winding the next. The insulation resistance 

 when finished, as tested on 100 volts, was over a megohm. The winding was 

 backed up at the ends by square ebonite washers which were fixed to the tube by 

 brass set screws. 



The Water-circulation. — A water-circulation was arranged for dissipating the 

 heat when large magnetising forces were used. A thin brass tube was chosen, so 

 that if it were slipped into the solenoid tube the annular space between them 

 would be about 2 mm. The smaller tube was centred by brass rings fitted at the 

 ends to the annular space between the tubes. The joint was made tight by solder- 



