4 Dr. W. S. Day on a Comparison of 



3. Manner of using the Thermometers. 



In comparing these various thermometers, it was necessary 

 to use them in the way in which they were originally used 

 when they were standardized. The Tonnelot thermometers 

 had to be used in the manner described by G-uillaume in his 

 ' Thermometrie de Precision' and in the printed notice that 

 came with the thermometers from the International Bureau. 

 Rowland's thermometers were used in a different way when 

 they were originally compared with the air-thermometer. 

 The chief points of difference are in the corrections for 

 internal and external pressure, and in the manner of taking 

 and using the zero point. The present method as used at 

 the International Bureau is to determine the pressure-coeffi- 

 cients, as they are called, by means of which any variation in 

 the external pressure from the standard 760 millim., due to 

 a change in barometric pressure, or to immersing the thermo- 

 meter in a liquid, and any internal hydrostatic pressure of 

 the mercury on the bulb, due to the thermometer not being 

 horizontal, can be taken account of and the proper corrections 

 applied. There are two pressure-coefficients, the external and 

 the internal. The external pressure-coefficient is generally 

 denoted by the letter (3 e , and gives the fraction of a degree 

 the reading of the thermometer is changed by a change in the 

 external pressure of 1 millim. of mercury. The internal 

 pressure- coefficient ft can be calculated for French hard 

 glass by adding the quantity O0000154 to the external 

 coefficient. This quantity depends on the elastic constants of 

 mercury and of glass. It is generally a small fraction, 10 per 

 cent, or less, of the whole pressure- coefficient. It differs, of 

 course, for different kinds of glass, but this difference is a still 

 smaller fraction of the whole pressure-coefficient. 



The zero point is always calculated from the zero taken 

 immediately after each reading of temperature, before the 

 zero has had time to change. 



Rowland's thermometers were originally used in quite a 

 different way. In his comparisons, the vertical position was 

 considered as the normal one, and the thermometers were 

 always used in that position. Changes in external pressure 

 were not taken account of except when they were caused by 

 the pressure of water in the comparison-tank. The ther- 

 mometers, when not in use, remained at a temperature of 18° 

 to 20° O.j the temperature of the room. The zero points 

 were taken at the beginning of a series of comparisons. In 

 taking the zeros, the thermometers were immersed in ice for 

 an hour or two before the reading was taken, and then the 



