H. A. Rowland— Magnetic Effect of Electric Convection. 35 
From this must be subtracted the effect of the opening in the 
center, for which the same formula will apply. 
The magnetic action of the excess at the edge may be caleu- 
lated on the supposition that that excess is concentrated in a 
circle of a little smaller diameter, C’, than the disc; 
NE ak 2—k? 
WAG: 
where ae at ia a and F(x) and K(#) are complete ellip- 
tic integrals of the second and first orders respectively. 
he determination of the potential was by means of the 
spark which Thomson has experimented on in absolute meas- 
ure. For sparks of length 7 between two surfaces nearly plane, 
we have on the centimeter, gram, second system, from Thom- 
son’s experiments, 
V—V'=117°5(7+ 0135), 
and for two balls of finite radius, we find, by considering the 
distribution on the two sheets of an hyperboloid of revolution, 
r V/r+1+1 
Wrpi 9 VSrpi-l 
where r is the ratio of the length of spark to diameter of balls 
and had in these experiments a value of about 8. In this case 
V—V'=109°6 (74-0135). (6) 
A battery of nine large jars, each 48-¢. m. high, contained 
the store of electricity supplied to the disc, and the difference 
of potential was determined before and after the experiment by 
charging a small jar and testing its length of spark. Two 
determinations were made before and two after each experi- 
ment, and the mean taken as representing the potential during 
the experiment. 
The velocity of the dise was kept constant by observing a 
governor. The number of revolutions was the same, nearly, 
as determined by the sizes of the pulleys or the sound of a 
beck siren attached to the axis of the disc; the secret o 
this agreement was that the driving cords were well supplied 
with rosin. The number of revolutions was 61: per second. 
In such a delicate experiment, the disturbing causes, such as 
the changes of the earth’s magnetism, the changing temperature 
of the room, &c., were so numerous that only on few days could 
numerical results be obtained, and even then the accuracy 
could not be great. The centimeter, gram, second system, was 
used 
V—V'=117'5 (J-+--0135) log. 
First Series. a=205, b=9-08, n=-697, D=110, H—-182 
nearly, B=1°68, 6=50, C=1055, N—61+, v=28,800,000,000;, 
n' =-0533, C’=10. 
