dU 



Prof. H. L. Callendar on a 



values so calculated to allow for the expansion of the bulb is 

 given by the formula 



Expansion correction to t = (c + bd)t(t — 100), 



where 6 stands for the absolute temperature (£ + 273). The 

 form of the correction is the same for the constant- volume 

 and for the constant-pressure type of thermometer. The 

 exact value of the correction may be affected to a slight ex- 

 tent by the method of calculation, and by the other correc- 

 tions, but the above formula is sufficiently approximate for 

 the purpose of estimating the effect of various assumptions 

 with regard to the expansion of the bulb. 



Table III. — Expansion Corrections of Gas-thermometers. 







Correction at 445° C. 



Correction at 1000°C. 



Material. 



Observer. 















c. 



b. 



Total. 



c. 



b. 



Total. 



Soda-glass. 



Callendar. 



4-70 



190 



6-60 







Hard-glass. 



Callendar. 

 Chappuis. 



327 

 3-60 



046 

 172 



373 



532 









Porcelain. 



Chappuis (1). 

 „ (2). 

 Bedford. 

 Holborn. 

 Deville. 



P38 

 138 

 1-63 

 2 04 



1% 



1-53 

 POO 

 036 



291 

 2-38 

 lfc9 

 2-04 

 P96 



o 



81 



8-1 



9-5 



1P9 



1P4 



O 



15-8 



103 



3-6 



23-9 

 18-4 

 131 

 119 

 1P4 



Plat.-Irid. 

 Platinum. 



Chappuis. 



Holborn. 



Callendar. 



407 

 o-85 

 402 



G-79 



53 



486 



3-85 

 4-55 



23-8 

 22-5 

 235 



8-2 

 5 V 5 



320 



225 

 290 



The correction is divided into two parts due to the co- 

 efficients c and b respectively. The first part increases 

 roughly as the square of the temperature, the second as the 

 cube. It is generally possible to determine c with much greater 

 accuracy than b, so that the uncertainty of the correction 

 increases considerably at high temperatures. The first two 

 determinations given in the table belong to the thermometers 

 used in 1890 and 1886 respectively, which agreed in giving 

 the value 444 0, 53 for the S. B.P. The hard-glass was a par- 

 ticularly hard specimen as a portion of it was successfully 

 used for a determination of the B.P. of cadmium at 756° C. 

 The measurements of the expansion were both made by the 

 same method as that adopted recently by Bedford for the 

 expansion of porcelain, and extended over the whole range 0° 



