156 PROCEEDINGS OF TIIE AMERICAN ACADEMY 



out stopping them. As cords and pulleys are used which may slip on 

 each other, it cannot possess much accuracy. I have devised a method 

 by cog-wheels which is more accurate, but which is better adapted for 

 use in the machine-shop than for scientific experimentation. 



But the most accurate method known to engineers for measuring 

 the work of an engine is that of White's friction brake, and on this I 

 have based my apparatus. Him was the first to use this principle in 

 determining the mechanical equivalent of heat. In his experiment a 

 horizontal axis was turned by a steam-engine. On the axis was a 

 pulley with a flat surface, on which rested a piece of bronze which was 

 to be heated by the friction. The moment of the force with which 

 the friction tended to turn tlie piece of bronze was measured, together 

 with the velocity of revolution. This experiment, which Ilirn calls a 

 balance de frotfemeut, was first constructed by him to test the quality 

 of oils used in the industrial arts. He experimented by passing a 

 current of water through the apparatus and observing the tempera- 

 ture of the water before and after passing through. He thus ob- 

 tained a rough approximation to Joule's equivalent. 



He afterwards constructed an apparatus consisting of two cylinders 

 about oO cm - in diameter and 100 om long, turning one within the other, 

 the annular space between which could be filled with water, or through 

 which a stream of water could be made to flow whose temperature 

 could be measured before and after. The work was measured by the 

 same method as before. 



But in neither of these methods does Him seem to have recognized 

 the principle of the work transmitted by a shaft being equal to the 

 moment of the force multiplied by the angle of rotation of the shaft. 

 In designing his apparatus, he evidently had in view the reproduction 

 in circular motion of the case of friction between two planes in linear 

 motion. 



Since I designed my apparatus, Puluj * has designed an instrument 

 to be worked by hand, and based on the principle used by Ilirn. He 

 places the revolving axis vertical, and the friction part consists of two 

 cones rubbing together. But no new principle is involved in his 

 apparatus further than in that used by Ilirn. f 



* Popg. Ann., clvii. 437. 



t Joule's latest results were published after this was written, and I was not 

 aware that lie had made this improvement until lately. The result of his 

 experiment, however, reached me soon after, and I have referred to it in the 

 paper, but I did not see the complete paper until much later. 



