the late Dr. Joule's Thermometers. 479 



researches which have been constructed on that basis. The 

 values of the scales of the thermometers A and B were 

 ascertained by plunging them along with the standard in large 

 volumes of water kept constantly at various temperatures. 

 The value of the scales of thermometer C was determined by 

 comparison with A. It was thus found that the number of 

 divisions corresponding to 1 = Fahr. in the thermometers A, 

 B, and C were 12*951, 9*829, and 11*647 respectively. And 

 since constant practice had enabled me to read off with the 

 naked eye to -J^- of a division, it followed that -g-o o °f a degree 

 Fahr. was an appreciable temperature.'" (Phil. Trans. 1850, 

 pt. i.; Collected Works, vol. i. p. 302.) 



On the Centigrade scale the figures given in the last 

 quotation would be 23'312, 17-692, 20*965. This allows us 

 to identify the second thermometer of the first quotation with 

 the one called A in all subsequent papers. 



" The thermometer used to indicate the temperature of the 

 calorimeter was the same which I employed in my former 

 experiments. Those designated A and D were calibrated 

 with great care. I have recently compared them together at 

 50 different temperatures between 32° and 80° Fahr., the 

 result being that, if the less sensitive was assumed to be 

 correct, the other, or A, nowhere appeared more than 0°*023 

 in error ; but taking averages for each consecutive 10°, this 

 error amounted to no more than 0°*008.' (Phil. Trans. 

 1878, part ii.: Collected Works, vol. i. p. 636.) 



Description of the Thermometers. 



The two thermometers which I had at my disposal were 

 those called A and D. The form and size of their bulb and 

 the width of the stem are shown in figs. 1 and 2, the former 

 representing in natural size the bulb and beginning of the 

 stem of the thermometer A, and the latter that of D. The 

 diameters of the two stems are 0*7 centim. (A) and 0*75 

 centim. (D). The length of the stems 87 centim. (A) and 

 86 centim. (D). The volumes of the bulbs maybe calculated 

 approximately from their shape, and are found to be 4*9 

 cub. centim. and 3*8 cub. centim. respectively. 



As the thermometer A was the one always employed in 

 calorimetric measurements, it is this instrument which is of 

 chief interest to us now. From the pressure coefficients of the 

 thermometer, we may approximately calculate the thickness of 

 the glass walls of the bulb, in the manner indicated bv 

 Gruillaume. The calculation can be carried out if the bulb is 

 cylindrical or spherical, and cannot strictlv be applied to such 



2K2 



