384 Dr. N. Ernest Dorsey on the Surf ace-Tension of 



which projects horizontally from the rod, and whose plane is 

 inclined to the rod at an angle of about 50° or more. The 

 lengths of the needle-point and of this short piece of spring 

 are such that when they are in place and the pendulum is at 

 rest they overlap by an amount just sufficient to ensure the 

 depressing of the needle-point at each complete vibration of 

 the pendulum. Both sides of this piece of spring are care- 

 fully polished, and the collar is so adjusted that the needle- 

 point comes just below the upper edge of the spring. The 

 electric circuit is completed through the long spring and the 

 mercury cup. Then, when the pendulum swings in one direction 

 the platinum point is depressed into the mercury closing the 

 circuit, and it immediately rises and breaks the circuit as the 

 pendulum swings on; on the return swing the platinum 

 point is slightly raised. If the damping vane works properly 

 the vibrations of the spring will be practically dead-beat, and 

 we shall have contact made once only during each complete 

 vibration of the pendulum. 



Dr. J. F. Mohler kindly compared the clock I used with 

 the astronomical one, and found that each single swing of 

 the pendulum corresponds to 1"000397 seconds ; the mean 

 solar day being the unit of time. Hence we are justified in 

 taking the complete period of the pendulum as two seconds. 



As any change in the position of the blades S 2 affects the 

 period of the fork, its frequency was often determined so 

 as to avoid errors due to careless handling. 



Amplitude of Waves. 

 We have seen that Lord Kelvin's equation applies to 

 waves of infinitely small amplitude, so it is well to consider 

 the amplitude of the waves used in this work. 1 have no 

 means of measuring its size directly, but a general idea of its 

 smallness may be obtained by considering the fact that 

 throughout the work the telescope was focussed for a point 

 one metre below the surface of the liquid. Since the beam 

 of incident light is parallel, this means that in the neighbour- 

 hood of a crest the radius of curvature is about two metres. 

 This, together with the fact that these waves are only 

 0*48 centim. long, shows that the amplitude must be extremely 

 small. We can also see that the amplitude is very small 

 from the effect of the meniscus upon the apparent length of 

 the waves at such a great distance as two centimetres from 

 the plate (p. 385). 



Effect of Viscosity. 



Before going further it is well to consider another point. 



