Royal Society. 315 



render the arrangements clear : and it is only necessary to add that 

 the whole is covered over^with an air- tight glass shade, 15 inches in 

 diameter and 16 inches high. 



3. The deflection of the galvanometer-needle, produced by heat- 

 ing the pile, is recorded by means of a small mirror attached to the 

 needle, which, as it moves, causes the reflected image of a line of 

 light to travel over a graduated scale. The galvanometer-needle is 

 rendered very nearly astatic by means of an auxiliary magnet ; and 

 this arrangement can be made so sensitive that if the temperature 

 of the disk, exposed to the cone as in the figure, were to rise l°Fahr., 

 this would be denoted by a change in the position of the line of light 

 equal to fifty divisions of the scale. 



4. In these experiments the disk was rotated rapidly for half a 

 minute, and a heating effect was, in consequence of this rotation, 

 recorded by the pile. The object of this paper is to investigate the 

 origin of this heating effect. 



5. In this investigation the authors prefer discussing the effect 

 produced on a metallic disk. The metal aluminium was chosen, 

 from its lightness, so as to diminish the weight upon the bearings as 

 much as possible. The reason for preferring a metallic disk is that 

 the heat produced in this case affects the whole substance of the 

 disk, and can thus be approximately measured. The disk of this 

 metal employed was ^th of an inch thick and 13 inches in diameter; 

 it weighed ten ounces, and in most of the experiments it was covered 

 with a coating of lampblack, applied by means of negative photo- 

 graphic varnish. In some of the experiments a plate of rock-salt, tightly 

 secured in a brass fitting, was screwed upon the mouth of the cone. 

 When this was done, a small piece of anhydrous baryta was placed 

 within the cone to keep the inner surface of the salt dry, and a dish 

 containing strong sulphuric acid was likewise placed in the receiver. 

 Indeed the latter was always used ; so that in the results obtained 

 the residual air may be considered as nearty dry, and the surface of 

 the disk, as well as that of the rock-salt, when this was used, nearly 

 free from moisture. 



Furthermore, in order to obviate the objection that the electric 

 currents which take place in a revolving metallic disk might alter 

 the zero of the galvanometer, the position of the line of light was 

 read before the motion began, and immediately after it ceased, the 

 difference being taken to denote the heating effect produced by the 

 rotation. The turning was made in this way: — As soon as full speed 

 was obtained, which might be about 10 seconds after beginning the 

 motion, a chronometer was noted, and the handle was turned at a 

 uniform rate for 30 seconds, and thereafter stopped as soon as pos- 

 sible. The most convenient speed in vacuo was found to be 20 revo- 

 lutions of the handle, or 2500 of the disk, in 30 seconds. It is believed 

 that the heating effect recorded may be considered as due to about 

 40 seconds at full speed. 



6. The thermometric value of the indication given by the galva- 

 nometer was found in this way: — The disk was removed from its attach- 

 ment and laid upon a mercury-bath of known temperature. It was 



