lO SCHMITZ, On a Method of Ice Calorimetry. 



vessel and on the balance-pan produced in some cases a 

 small error ; I found that this might amount to a few 

 centigrammes in a protracted weighing, but it was easily 

 reduced to one or two centigrammes by arranging before- 

 hand for a quick weighing. The error due to convection 

 currents was found by experiment to be very small. 



I have not applied any correction for the weight of air 

 displaced by the ice. Such a correction (one part in one 

 thousand) would be rather greater than the correction for 

 deposition of dew in my experiments and of opposite 

 sign ; for the present purpose the two corrections may be 

 taken as neutralising one another. 



Temperature of the Liquid Air. 

 To determine the temperature of the liquid air I made 

 a platinum thermometer with exposed coil. My thanks 

 are due to Mr. J. E. Petavel for information as to details 

 of construction. The readings were taken with an Elliott 

 box (previously standardised) and a single Leclanche cell. 

 The resistances at o'^C. and ioo°C. were determined with 

 great care, and with various modifications of method. In 

 the absence of a supply of liquid oxygen the resistance at 

 — 1 82*5 '^C. (taken as the temperature of liquid oxygen) was 

 determined by placing the instrument in a narrow test- 

 tube from which the last remainder of a quantity of liquid 

 air was allowed to evaporate ; the steady resistances 

 obtained in a number of independent experiments by 

 this method agreed very closely. 8 (of Callendar's well- 

 known formula) was found to be 2'47 (the last figure of 

 doubtful significance). This value may be compared with 

 the following numbers given by Professor Dewar* for 

 two platinum thermometers : — 2*5 797, 2'6'/6'/. The wire 

 was of pure metal, 36 S.W.G., obtained from Messrs. 

 Johnson, Matthey, & Co. 



* Proc. Roy. Soc, Vol. 68, p. ^fil, 1901. 



