216 J. A. POLLUCK. 
are perfectly symmetrical round this point, is taken as the 
distance which makes the condenser circuit in tune with 
the rectangle. 
This operation of tuning the circuits is a tedious one; it 
has involved in some cases the comparison of fifteen com- 
plete curves drawn from observations obtained as just 
described. Once the circuits are tuned. the series of obser- 
vations necessary for a determination of the length of a 
third circuit when in tune with the rectangle, need only be 
repeated until the chance of a fortuitous accordance of the 
observations is eliminated. 
Typical examples of the curves obtained are given in 
Fig. 5. The ordinates represent the difference of reading 
in cms. of the deflections caused by the detector before and 
after demagnetisation. ‘They are proportional to the mag- 
nitude of the disturbances in the third circuit. Abscissee 
represent the length of the third circuit, observations being 
usually taken for successive lengths differing by 20 cms. 
As the demagnetisation of the detector depends, cceteris 
paribus, on the length of the spark, such curves are only 
wholly comparable when they represent observation taken 
with the same spark gap. 
In Fig. 5, curve 1 is a plot of observations taken in con- 
nection with one of the straight wires using the solenoidal 
detector. Ourves 2, 3, 4, refer to the case of one of the 
open circles. They are the plots of series of observations, 
without alteration of spark gap, taken with condenser plate 
distances successively increased by 0°1 cm., the solenoidal 
detector being used. For tuning the condenser and rect- 
angular circuits thirteen such curves were obtained, the 
series being extended and repeated three times to avoid 
any chance accordance of the observations. Curve 5 is a 
plot of observations with a closed ellipse using the longi- 
tudinal detector (see infra). In this case the observations 
