the late Dr. Joule's Thermometers. 



485 



probably unique in having been watched for a period of over 

 50 years. Joule furnishes us with data showing the gradual 

 rise of the zero from April 1844 to March 4, 1873 (Collected 

 Works, vol. i. p. 558). In the communication he presented 

 to the Manchester Literary and Philosophical Society, he 

 refers the rise to the first observed zero without giving the 

 actual readings, so that this short paper dees not allow us to 

 judge how far the present zero is above that of 1 873. For- 

 tunately, we can indirectly supply the deficiency. For in his 

 comparison with Rowland's thermometer (Proc. Amer. Acad, 

 vol. xvi. p. 38) the reading of the zero is stated to be 22*62. 

 Rowland's thermometer was sent to Joule in the summer of 

 1879, and the results were communicated to the American 

 Academy in March 1880. The comparison must have been 

 made at an intermediate date, and can therefore be identified, 

 for in the above-mentioned communication only one com- 

 parison is mentioned between January 1877 and December 

 1882, and that one in November 1879, when the zero stood 

 12*92 divisions above that of the first observation. Taking 

 the original zero to be 9*7, the complete series now is as 

 follows : — 



Table III. 



Date. 



Zero. 



Date. 



Zero. 



April 1844 

 February 1846... 

 January 1848 ... 

 i April 1848 . 

 February 1853... 

 April 1856 . , 

 December 1860... 

 March 1867 . 

 February 1870 ... 



9*7 

 15 2 

 16-3 

 16-6 

 18-5 

 192 

 20-8 

 21-5 

 21-8 



February 1873... 

 January 1877 ... 

 November 1879... 

 December 1882... 



:22 



22-41 



22-62 



2296 



April 1892 



23-36 (17°) 

 23-31 (17°-6V 

 23-35 (17°) ' 



April 1893 



June 1894 





I have taken a considerable number of readings of the 

 zero-point since the beginning of the year 1892. They vary 

 of course with the temperature to which the thermometer was 

 exposed. The determinations were made in an apparatus 

 similar to that described by Guillaume, the thermometer being- 

 immersed in a mixture of scraped ice and distilled water. Great 

 care must be exercised in the readings, for, owing to the large 

 size of the bulb and the long time taken by the thermometer 

 to reach a steady state, the results are easily vitiated by an 

 accumulation of water. No correction was made for the 

 pressure due to the surrounding mixture, as probably Joule 

 took no account of that pressure, nor would the correction be 

 significant for our purpose. The readings are arranged \r\ 



