PERIODS OF.ELECTRICAL VIBRATIONS. 217 
Fig. 5. 
near the maximum were taken with ellipses differing in 
perimeter by 10 cms. 
The maximum is much more definitely marked in the case 
of open circles than in circuits of other shapes and here a 
want of tune between the condenser and rectangular circuits 
seems to have less effect in altering the character of the 
curve of the observations than in other cases. It has been 
found a more difficult matter to get a series of accordant 
observations with closed circuits than with open ones. 
The Sparks in the Condenser Circuit.—Practically the 
only difficulty in connection with the determination of 
numerical relationships in the case of Hertzian waves of 
small wave length lies with the spark which discharges the 
condenser. Many forms of detector are completely reliable. 
On the other hand the character of the sparks is so variable 
and the conditions which determine it, at present so obscure, 
that chance seems to enter largely into these investigations. 
Many spark knobs have to be tried in some cases before 
the observations become quite accordant. To obtain a 
O—Dec. 2, 1903. 
