THERMAL CONDUCTIVITIES OF SOME ELECTRICAL INSULATORS. 43 J 



iii-;it h by immersion for about I centim. in cold water. As the liquid cooled and 

 solidified in tbe lower part of the tul>e, more liquid was supplied at the top from ft 

 heated test tube. This was continued till the whole of the brass tube was filled 

 with the solid material, to a little above the upper edge. The top was then heated 

 and brought down on to the material, melting it as it came so that the two were in 

 good thermal contact. 



The apparatus was then placed centrally in the Dewar tube, as shown in fig. 2, the 

 circuits made, and the difference of the resistance of the two platinum spirals and 

 the resistance of one of them were then measured. If these were found to be 

 normal, a current of I or 2 amperes was sent through one of the heating coils, and 

 the two resistance measurements repeated after 5 and 10 minutes' intervals. If 

 the two sets of olervations were equal, or nearly so, the last observations were 

 taken to represent the temperature for the steady state. During the intervals 

 l)etween the readings, any small change in the heating current was corrected by a 

 re-adjustment of the external resistances. The heating current was then switched 

 off the first heating coil and sent through the second, and the observations repeated. 

 The mean of the two values of the difference of resistance of the platinum spiral, 

 and the mean of the resistances of the hotter spiral in each case, were then taken. 

 These means eliminate from the result any want of symmetry in the heating and 

 temperature measuring coils, and in their positions in the brass tube. 



If any unusual difference between the observations with the heating currents 

 through the two coils was observed, a third set of observations with the heating 

 current again through the first coil was taken, and the mean result for these two 

 olmervations was combined with the result with the heating current through the 

 other coil. 



These tests having shown that the apparatus was in working order, it was removed 

 from the Dewar tube, which was then half filled with liquid air, and the apparatus 

 lowered very slowly into it, so that at first not more than I centim. of the brass tul>e 

 would IH? immersed, till a considerable length of the tube and contents had been 

 cooled down to a low temperature. Then the tube would be lowered still further, 

 and so on till completely immersed. 



When this was the case, a thick plug of cotton wool was placed in the top of the 

 Dewar tube. 



During the cooling process an occasional test of resistance would be made to see 

 how the cooling was proceeding. 



When the tube was completely immersed in the liquid air, the heating current was 

 adjusted to 1 or 2 amperes, as the case might be, and the observations taken as 

 described above. 



When sufficient observations at the temperature of the liquid air had been made, 

 the Dewar tube was tilted in such a way that the liquid air would run out without 

 the apparatus being disturbed. 



