A ci.ocK-coM Roi.i.r.i) ri'Nrxa i-<u<k i.vi 



freiiia-ncies luuiiiK ratios as IurIi as UK) to 1 may be readily coin- 

 partxl. For ratios of the order of 100 to 1 the lower frefiuency is 

 preferably stepptxi up to a high voltage to give an c<iuivalent deflec- 

 tion of as much as 25 centimeters, thus giving a spacing between 

 cv-cles for the high fre<.|ucncy of approximately 0.5 centimeter. Of 

 course, the whole 25 centimeter deflection is not shown on the screen 

 but this is unnecessarj'. The value of the ratio cannot be at once 

 determinetl by this means, there being no appreciable dilTerence 

 between the figure for a ratio of 100 to 1 and 99 to 1, but this ratio 

 may be readily determined liy aimparing each frequency separately 

 with an intermediate frwiucncy such as 500 cycles. 



Having determined the ratio between the high and low frequencies, 

 it is possible, by drawing a reference line across the screen, to de- 

 termine whether or not they are keeping step with one another. Thus 

 for a comparison of 50 cycles against 5000 cycles, if we get a motion 

 of 2 waves in 10 seconds, this represents a deviation from exact syn- 

 chronism of 2 parts in 50,000. 



Comparisons made in this way between the 50-cycle fork and a 

 vacuum tube osc:illator giving a constant frequency of .5000 cycles 

 show no deviation in the mean period of the fork greater than 1 part 

 in 50,000 for observations extending over several minutes. If devia- 

 tions greater than this were observed, they might equally be ascribed 

 to the auxiliary- oscillator but the fact that they do not occur means 

 either that the fork is constant to better than 1 part in 50,000 or that 

 both frequencies varv- in e.\actly the same way which is ver>" im- 

 probable. 



The above method of comparison does not require a sine wave of 

 current from the fork. In fact it has been found advantageous to 

 have a somewhat distorted wave since an unsymmetrical figure on 

 the luminous screen of the tube is more easily observed. This is 

 due to the fact that one-half of the figure moves across the screen in 

 one direction while the other half moves in the opposite direction. 

 In order not to confuse one half with the other, it is highly desirable 

 that they be dissimilar in shape and this is accomplished by using a 

 distorted wave as the lower frequency. Sufficient distortion is 

 secured by mounting an ordinary telephone receiver in close prox- 

 imity to one prong of the fork as shown in Fig. 1 and amplifying the 

 e.m.f. thus obtained as much as necessary- to obtain the desired voltage. 



By means of the simple control system described above, it has 

 been possible to obtain a fundamental frequency so free from fluc- 

 tuations as to be constant over short or long periods of time to ap- 

 proximately one part in 100,000. 



