226 Dr. D. K, Morris on the Magnetic Properties 



following day the ring was again heated to this temperature, 

 it being intended to take a series of observations during 

 cooling. The series had, however, to be postponed, owing to 

 an unfortunate occurrence — the sealing-wax over the cork 

 had slightly softened during the prolonged heating, and the 

 excess of outside pressure began to drive the cork slowly into 

 the jar, rendering it no longer air-tight. The heating current 

 was immediately shut off; but, when cold, the iron core was 

 found to have been partially oxidized by the inrush of air, 

 reducing its section about 4 per cent. To prevent a repetition 

 of this occurrence, an indiarubber tube containing flowing 

 water was arranged to keep the neck of the jar cool. 



During the second absorption of the oxygen, the iron-wire 

 coil, which had already been used once for this purpose, burnt 

 out. But it was thought that the absorption might be com- 

 plete, and the experiments were proceeded with. After heating 

 the ring, therefore, a third time to 840°, a series of measure- 

 ments was made while the temperature was falling. 



The ring was then still further annealed by heating to 1150° ; 

 and, during cooling, a set of observations like the previous 

 was taken. On their completion, this ring was also taken to 

 pieces that the cross-section of the core might be re-determined 

 and the constants of the thermometer-wire checked. The 

 last absorption of the oxygen does not seem to have been at 

 all complete, for during the last heating the section diminished 

 by 6 per cent. 



Thus, in these experiments, though the method of absorbing 

 the oxygen was shown to be efficient, yet troubles due to 

 oxidation were by no means absent. Irrespective of the 

 diminution of the sectional area of the iron, the presence of 

 oxide in the core might affect the results in a variety of 

 ways: — (1) by the shunting effect due to its conductivity, or 

 (2) by impairing the quality of the insulation, both tending 

 to diminish the apparent resistance of the iron ; and (3) by 

 having an appreciable permeability, affecting the magnetic 

 measurements. 



(1) The shunting effect is quite negligible. (2) Though 

 the inner surface of the mica was altered by contact with the 

 oxide, the insulation was not seriously injured. (3) With 

 regard to the third possibility : — In small fields, such as those 

 used in the present experiments, the magnetic qualities of 

 soft iron are at ordinaiy temperatures so very much more 

 intense than even those of the magnetic oxide of iron, that 

 no perceptible alteration of the ballistic galvanometer throws 

 could arise from any oxide present. At temperatures above 

 the critical point, the experiments of M. Curie, already cited, 



