COMPARISON OF PLATINUM AND GAS THKRM< >\li: IKKS. 63 



[Paragraph added December 1, 1899. The measurements by which we attempted 

 to determine the scale of the platinum thermometers may be divided into four groups, 

 in which different instruments and means of heating were employed, and in which 

 the precision varied from group to group. 



These are 



(1) Comparisons in water between and 50 of platinum thermometers K.8 



and K.9 with the four principal mercury standards of the Bureau. 



(2) Comparisons of K.8 and K.9 in an oil bath at temperatures between 80 



and 200, with a constant volume nitrogen thermometer, the initial 

 pressure of the gas being 793 millims. of mercury. 



(3) Comparisons of thermometers K.8 and K.9 between 250 and 460 in a bath 



formed of a mixture of nitrates of potash and soda, with the nitrogen 

 thermometer, the initial pressure being 529 millims. 



(4) Comparison of thermometer K.2 with the same nitrogen thermometer in the 



same bath Ixjtween the temperatures 424 and 586, tin- initial pressure 

 being 392 milling. 



As the sensibility of the gas thermometer varies according to the initial pressure, 

 it is evident that the same precision cannot be attained in the different series. The 

 construction of our instrument was such that the highest measurable pressure was 

 about 1400 millims.] 



XXIV. GENERAL CONSIDERATIONS ON THE GAS THERMOMETER. 



In accordance with the decision of the International Committee of Weights and 

 Measures,* the provisionally accepted normal scale of temperature is that of (In- 

 constant-volume hydrogen thermometer. The employment of hydrogen for our work 

 seemed therefore advisable, and before proceeding to the actual comparisons, we made 

 a number of trials of the hydrogen thermometer between 100 and 200. Up to 

 trni|>eratures about 180 these experiments gave fairly good results, hut we noticed 

 that prolonged heating alx>ve 180 was generally followed by a diminution of the gas 

 contained in the thermometer reservoir. This diminution, though small, being 

 regularly reproduced after each prolonged heating, might become serious at higher 

 temperatures. Some S|K rial measurements, made on a known quantity of hydrogen 

 enclosed in a capillary "I " verre dur" of I square millim. cross section, and exposed 



* The resolution fixing this was parsed l.y the International Committee on October 15, 1887, and is M 

 follows : 



Th.u the International Committee of Weights and Measures adopt as the Normal Thermometric Scale 

 for tin- International Service of Weights and Measures, the centigrade scale of the Hydrogen Thermometer 

 having ,-i.s lixrd ]x>inu the temperature of melting ice (0), ami that of the vaj>our of distilled water in 

 ebullition (100 ) under the normal atmospheric pressure ; the hydrogen being taken under the manometric 

 initial pressure of one metre of mercury, i.r., at -Wy T3158 of the atmospheric pressure." 



