472 Proceedings of Royal Society of Edinburgh. [sess. 
surface. According to our former experiments* the quantity 
flowing from either the uranium or from the surrounding metal per 
sq. cm. of its surface increases but little with increased voltage 
when this exceeds 5 or 10 volts per cm. Now, when the greatest 
diameter of the uranium is small in comparison with distances to 
the outer metal surface, the voltage per cm. is much greater along 
the lines of force near the uranium surface than near their outer 
ends on the surrounding metal. Hence the rate of discharge of 
electricity into the air from the uranium will cease to increase 
sensibly with the difference of potential between the uranium and 
the surrounding metal, while the rate of discharge of the opposite 
electricity from the large surrounding metal surface is still notably 
increasing. Hence if the dimensions and shapes of the uranium 
and of the surrounding metallic surface are such that for small 
voltages, such as 10 or 20 volts of difference between the 
uranium and the surrounding metal, the electricity lodged in 
the air by discharge from the uranium preponderates over that 
discharged from the surrounding metal, the excess must come 
to a maximum and diminish, possibly even down to zero, with 
greater differences of potential : and at potential differences 
still greater the electricity lodged in the air from the outer metal 
may preponderate, and the electricity in the air drawn off and 
given to the filter be of opposite sign to that of the uranium which 
was found with the lower voltages : provided the configurations 
are such , and the voltages are so moderate that disruptive discharge 
does not intervene to any practically disturbing extent. 
* Kelvin, Beattie, and Smolan, Proc. R.S.E., 1897. 
