each fragment being about 3 mm. square and 1 mm. thick. Four of these fragments were affixed to each metal plate with a touch of shellac at the four corners and one fixed in the middle. The seventeen plates were then piled one on the other, the glass fragments acting as separators, and the alternate plates were connected together by wires soldered to each series. A. metal clamp kept all the plates in position. The condenser so formed consisted of seventeen plates, eight being the positive, and nine the negative surfaces. The glass distance pieces had a total surface of very nearly 1 per cent, of the total opposed surface of the plates. The condenser so formed had a capacity of 0-001031 of a microfarad when gaseous air at 15° C. and normal pressure formed the dielectric.
If such a condenser having a capacity C' is charged to a potential Y and then discharged n times in succession into a larger reservoir condenser of capacity C, it is easy to show that at the end of the n successive charges the quantity Q contained in the large condenser is given by the series
Q = C 'V ( m to2-f-to3 . . . . to"), where m Hence we have Q = C'V‘n (l — to’*). 1—to
The capacity C' of the small aluminium condenser may he considered to he made \ip of two parts; a part which is changed when liquid oxygen is substituted for gaseous oxygen or air on immersing the condenser, and which thereby becomes increased. If K is the dielectric constant of liquid oxygen, referred to that of gaseous oxygen at —182° C. as unity; and if c is the capacity of this variable part of the condenser when the dielectric is gaseous oxygeD, then Kc is its capacity when liquid oxygen is substituted for the gaseous oxygen at the same temperature.
In the next place there is a small part of the whole capacity due to the glass separators. These, as a whole, have a surface very nearly equal to 1 per cent, of the whole surface of the metal plates, and a dielectric constant, as shown below, when cooled to —182° C., of 5'0. Hence it follows that that part of the whole capacity of the condenser which is due to the glass separators, may be represented very nearly by 5c/100.
This part of the capacity remains practically constant whether the condenser is lifted out of the liquid oxygen into the cold gaseous oxygen lying above it, and which is at nearly the same temperature, or put into it, as long as the condenser is very nearly at the same temperature in the two conditions.
