310 



Royal Society. 



dimensions of the agitator are so arranged, that no part of it can 

 possibly touch the thermometers when in operation. The vase, 

 containing -svater, the stand -with thermometers, and the agitator, 

 is mounted upon a wooden revolving stand. The depth of water in 

 the vase is always sufficient to include the whole of the column of 

 mercury, the scales being observed through the water. In taking 

 the observations, the observer, after agitating the water briskly for 

 some time, turns the revolving stand till each thermometer is brought 

 successively opposite to his eye, reading off the scales as quickly as 

 possible to an assistant, who writes down the numbers. Proceeding 

 in this way, I find that six thermometers can be read off and recorded 

 easily in 20 seconds. It is of course desirable to make more than 

 one set of readings for each temperature ; and in order to avoid as 

 much as possible the changes which may occur during the reading 

 off, it is well to reverse the order of observing the instruments, that 

 is, to read them alternately in the order one to six, and six to one. 



The following table contains the results of comjjarisons of six 

 thermometers, and will show the accuracy which may be obtained 

 by the method of comparison just described ; it will also exhibit the 

 accordance in the indications of instruments graduated according to 

 Regnault's process. Each result is the mean of six comparisons. 

 No optical assistance was used in reading off the scales. The free- 

 zing-points of all the instruments were determined on the same day, 

 after the comparisons were made. 



I 



