596 Prof. C. Barus on the 



ments under proper adjustment * to over 18°. They were 

 thus strikingly visible across a large room even without an 

 attached index. It occurred to me that such an arrangement 

 could be made to possess considerable didactic and possibly 

 practical interest, being not only a capital illustration of the 

 principle of resonance and of the principle of areas, but 

 lending itself at once to a more serious study of the inter- 

 ference of two nearly synchronous vibrations with each other. 

 A very perfect system of two degrees of freedom undergoing 

 f jrced vibration is presented for inspection : for the watch is 

 a recording instrument, and whatever period supervenes 

 under given conditions is found from the dial in comparison 

 with a chronometer, also a recording instrument. 



"With regard to the motion one may infer in a partial way 

 that the inertia of the balance-wheel will be apparently 

 increased or diminished according as the period of the watch 

 is above or below the normal period of the former (about 

 I second). While the balance-wheel, apart from friction, is 

 a simple harmonic pendulum fr-r all excursions, the watch is 

 not rigorously so, but shows a longer or shorter natural period, 

 depending on greater or smaller elongations within the limits 

 of observation. This is an additional peculiarity giving further 

 scope to the instrument. 



On adding a suitable ballast at b (fig. 4), beats were natur- 

 ally obtained often striking in character. But a singular 

 feature was observed in these experiments : while beating 

 was marked at the outset, no adjustment was found which 

 made these beats permanent. Th^y invariably vanished 

 gradually in the lapse of time, usually into the uniform 

 harmonic motion of smaller amplitude and period. These 

 and other circumstances induced me to follow the pheno- 

 menon carefully, and in the following table I have recorded 

 a selection of the results obtained in so far as they deserve 

 notice. 



The watch was mounted on a steel knife-edge and swung 

 from hard steel plates (an old horseshoe-magnet) . It was 

 removed once in 2A hours to be rewound, though sometimes 

 36 hours were allowed to elapse. Comparisons were made 

 with a good chronometer in the morning and evening to 

 seconds, this being much greater accuracy than the watch 

 warranted. 



(2) Since the centre of gravity of the watch is approxi- 

 mately determinable by trial, and its period known, the moment 



* Much larger angles could have been obtained had the adjustment 

 permitted. Apart from this and friction &c. there is no near limit. 



