Critical and Boiling Temperatures of Hydrogen. 203 



only between those limits in which it has been compared with 

 the hydrogen thermometer ; every extrapolation would lead 

 to quite false results. 



After failing in my experiments with the thermo-electric 

 junction, I resolved to measure the critical temperature of 

 hydrogen by means of a platinum thermometer based on the 

 variation of the resistance of a very thin platinum wire with 

 the temperature. This method for measuring very low tem- 

 peratures was proposed by Messrs. Cailletet and Collardeau"*, 

 afterwards also by Dr. Guillaumef, but practically it was used 

 for the first time by Prof. WitkowskiJ. His researches 

 proved that rhe platinum thermometer is, after the hydrogen 

 thermometer, the most suitable for measuring low tempera- 

 tures. It is true that it demands an accurate comparison 

 with the hydrogen thermometer, because the variation of the 

 electrical resistance of platinum is not directly proportional 

 to the variation of the temperature ; but between those limits 

 in which it was compared with the hydrogen thermometer 

 it may be used to determine low temperatures very rapidly 

 and accurately. And considering that the curve which 

 represents the relation between the temperatures measured 

 with hydrogen and the electrical resistance of platinum is 

 very nearly a straight line, we may draw the conclusion that 

 by measuring very low temperatures by means of a platinum 

 resistance-thermometer we may venture on an extrapolation 

 to a moderate extent, and that the error which results from 

 the imperfect agreement of the variation of the resistance to 

 the variation of the temperature cannot surpass 0°*5 to 1°. 



The platinum thermometer of such a form as given by 

 Prof. Witkowski is suitable for many practical purposes, but 

 it does not follow variations of temperature with such a 

 degree of quickness as is indispensable for the experiments I 

 had to perform. The reason of ihe insufficient rapidity of 

 Prof. Witkowski's thermometer is the not very thin platinum 

 wire (0*06 millim. diameter), which was insulated with silk, 

 and hermetically enclosed between two concentric copper 

 tubes made of very thin copper foil. 



To give my thermometer the greatest possible sensitiveness 

 I endeavoured to remove all that could have any contrary 

 influence upon it. So, for instance, I used a much thinner 

 wire of chemically pure platinum, whose diameter was only 



* Journal de Physique, 1888. 



f Archives d. Sc. phys. et natur. Geneve ; also L 1 Industrie electrique, i. 

 p. 78. 



X Rozprawy (Transactions) of the Cracow Academy [2] iii. p. 380; 

 Bull. Intern, of the Cracow Academy, 1891. 



P'2 



