4384 H. A. Rowland—Absolute Unit of Electrical Resistance. 
n 7°8 seconds, but the time of vibration was too short and 
se needle was constructed vibrating in 11°5 seconds, which 
was a sufficiently long period to be used successfully after 
practice. 
There seems to be no error introduced by the time taken to 
reverse the commutator in the method of recoil, seeing that the 
breaking of the current stops the needle and the making starts 
it in the opposite direction. As the time was only a fraction 
of a second the error is minute iu any case. 
hile the current is broken in the reversal, the battery may 
recuperate a little and there is also some action from the extra 
current, but there seems to be no doubt that long before the 
four or six seconds which the needle takes to reach its greatest 
elongation everything has again settled to its normal condition 
and the current resumes its original strength. Hence the error 
from these sources may be considered as vanishingly small. 
Some experiments were made by simply breaking cscs current 
and they gave the same result as by reversal. 
The ovewieg is the order of observations corresponding to 
each experim 
“$3 he ene of —— of needle was observed. 
. The = was around the circle, F, so as to 
sa B and a. Sirmnitanenaa readings were taken at the 
two galvanometers. The commutator at the tangent galvan- 
ometer was then reversed and readings again taken. After 
that the ea paren to the circle was reversed and the operation 
s gave four readings for the circle and eight for 
the tangent arid claire as both ends of the needle were read. 
n some cases these were pore ey six and twelve respect- 
ively. This operation was repeated three times with currents 
of different eR coristitubing three observations each of 4 
and fp. T nate any action due to the induction coils, they 
were sometimes korisented in one way and sometimes in the 
opposite w 
BG. 
