270 Prof. Townsend and Mr. Morrell on Electric 



be noticed that the quantity ??u<; 2 /(w 1 — w 2 ) should be a whole 

 number, and it may be taken as the nearest integer to the 

 value obtained experimentally. 



For example, in one experiment in which thirty harmonics 

 of the long- wave generator were counted in changing the 

 wave-length of the short-wave generator from 12'12 to 

 7*996 metres the quantity miv 2 j{wi — w^), as found experi- 

 mentally, is 58*1, and neglecting the figure in the decimal 

 place as being due to experimental error in measuring the 

 difference (w l — w 2 ), the fundamental wave-length of the 

 long-wave generator is 



VV = 12*12x58 = 703 metres. 



In this case the wave-length W as indicated directly by the 

 wave-meter was 706 metres. 



The wave-length W may also be measured when the short- 

 wave generator is adjusted to a fixed wave-length w { , by 

 counting the number of harmonics n that come into tune 

 when W is increased continuously from a value Wi to 2Wj. 

 The factor n is thus known, and Wi = ww*i. In this method 

 W is adjusted by the variable condenser in the long-wave 

 oscillator, and it is necessary to have a means of determining 

 when W is changed from any wave-length Wj to 2Wj. For 

 this purpose an oscillatory circuit consisting of a fixed con- 

 denser and a variometer in which the two coils are of equal 

 self-induction was used. This circuit with the coils in parallel 

 is brought into tune with the oscillation Wi, and when the 

 coils are changed from parallel to series the variometer circuit 

 is again in tune when W attains the value 2W 2 . 



This method is useful in determining a series of waves 

 which are multiples of a simple standard length such as 

 10 metres. Wave-lengths up to 1000 metres as measured 

 by this method were found to be in agreement with those 

 given by the wave-meter, which was calibrated by the 

 National Physical Laboratory. 



6. In this method of measuring long waves it might be 

 supposed that a difficulty would arise if short-waved free 

 oscillations were maintained in the inductance coil of the 

 generator as well as the higher harmonics of the principal 

 tree oscillation. It is well known, for example, that a coil, 

 even when connected to a condenser, has several free modes 

 of oscillation with wave-lengths much shorter than that of 

 the principal free oscillation determined by the product 

 of the self-induction of the coil and the capacity of the 

 condenser. These free oscillations are of the type discussed 

 for a long solenoid in sections 7 and 8, and mav occur in any 



