March, 1918] A Lecher System — Experimental 



151 



for the bridge curves the check receiver helped to eUminate 

 oscillator variation, it proved wholly useless for spark-gap 

 curves. The check receiver was placed, after Blake and Sheard, 

 near the oscillator spark-gap, but in general the maximum for 

 the two receivers did not occur at nearly the same spark-gap 

 length, hence neither curve could be used to correct the other 

 except by an approximation method of some sort. Hence we 

 abandoned its further use. Accordingly, the only safe way to 

 make a direct comparison of the curves of Figure 3 was to get 

 some one tone, say the fundamental, to a common basis. 



fis 



S 7 9 ^f 



"ill 



Figure 5. 



Before describing the manner of reducing our corrections to 

 a common basis it is well to point out first, however, that the 

 optimum spark-gap length varies with the frequency-number 

 just as Blake and Sheard found (see their Figure 12). Curve 

 IV of Figure 5 summarizes roughly this relation as the mean 

 value for all the ys used. The optimum spark-gap length 

 for a given tone is not independent of y, however, as Curve V 

 of that figure shows for the fundamental. It is clear from 

 Curve IV, that the optimum spark-gap length decreases with 



