362 Dr. T. Ewan and Mr. Haldane Gee on 



the same in both the glass and the copper apparatus. The 

 difference at lower temperatures is probably due to the fact 

 that water does not boil regularly in a glass vessel under 

 diminished pressure. This irregular boiling gives rise to 

 sudden alterations in the pressure, and to overheating of 

 the steam. 



In making the measurements of temperature by the 

 platinum thermometer, the greatest error was (as has 

 already been mentioned) in the correction for the tempera- 

 ture of the resistance box. To eliminate this as much as 

 possible, we took readings of the resistance at 0° and 100° 

 before each set of observations, and then kept the tem- 

 perature of the coils as constant as possible, while the 

 readings at intermediate temperatures were made. 



In determining the zero point, the platinum thermometer 

 was left in the ice for at least an hour, and the current only 

 closed momentarily (the first swing of the galvanometer 

 needle being read), in order to avoid heating the wire. 



The current was also reversed in order to eliminate 

 thermal effects. No correction was made for the resistance 

 of the copper leads. 



The zero point was found at different times. 



Resistance at o\ 



Mean. 



Differenc. 

 from 

 Mean. 



10-6388 

 10*6401 

 10*6343 

 10-6360 



10*6372 



4- -0016 

 + '0029 



- -0030 



— '0012 



It is possible that the alteration in the resistance is due 

 to slight strains in the wire, as the protecting glass tube was 

 not rigidly attached to the wooden part of the thermometer. 



The following are the determinations of the resistance 

 at 100° in the order in which they were made, and it is 

 curious that they increase and decrease regularly. 



