and Electrical Resistance of Iron. 225 



The magnetizing current was measured with a tangent 

 galvanometer, whose indications were compared with those of 

 a standard Weston ammeter. The magnetizing force was 

 assumed to be uniform in intensity throughout that part of 

 the section of the ring occupied by the iron ; which assump- 

 tion is, in the present case, but approximately true, though 

 the oblong cross-section favours the uniformity. 



Description of the Experiments. 



The iron cores in their initial condition may be regarded 

 as unannealed ; for the original annealing was destroyed 

 during their preparation. 



To study the effects of annealing, measurements were taken 

 during the first heating of specimen A. With this exception, 

 however, all the experiments recorded in this paper were 

 made during the cooling from some definite temperature at 

 which the iron had been thoroughly annealed, that all effects 

 other than those due directly to change of temperatures might 

 be eliminated. The duration of each of the long sets of 

 experiments was about 20 hours. 



The ring containing specimen A was first heated slowly 

 up to 1050°, stopping at various stages on the way up for 

 observations, and then allowed to cool, when another series of 

 observations was made. It was afterwards heated to 920° 

 to find whether the annealing was complete. 



The ring was then taken to pieces, that the cross-section of 

 its iron and the constants of the thermometer-wire might be 

 re-determined. 



The arrangements for absorbing the oxygen were with this 

 ring not very perfect. The coil for this purpose had but one 

 iron wire instead of two in parallel ; and, becoming locally 

 oxidized, the wire fused before it had absorbed all the oxygen. 

 The cross-section of the iron core on removal from the ring 

 and on detaching the oxide was found to have diminished 

 by as much as 8 per cent. Since, however, no change of 

 section occurred (as deduced from the constancy of the 

 resistance and permeability) during the concluding heating 

 to 920°, it may safely be assumed that the oxidation took 

 place previous to the cooling from 1050° ; and that the section 

 during subsequent experiments was the same as that measured 

 at their close. 



The ring-magnet, of which specimen B was the core, was, 

 after preliminary magnetic measurements, heated at once to 

 840°, a little above the critical point ; at this temperature it 

 was kept for about two hours and then allowed to cool. The 



