438 PROCEEDINGS OF THE AMERICAN ACADEMY. 



A large storage battery exhibits in an interesting manner the time 

 necessary to change condensers to a definite potential. When the inner 

 coating of a Leyden jar of about 10,000 electrostatic units is connected 

 with one pole of the battery, and the other rests upon a wooden table upon 

 which the Leyden jar is placed, this jar will discharge through a small 

 spark gap once in about ten seconds. When this latter terminal is con- 

 nected directly with the outer coating of the jar, the rapidity of discharge 

 increases to at least one hundred a second, and appears to the eye to be 

 continuous. The battery acts like a reservoir, and continually supplies 

 the loss in the jar. I was interested to ascertain whether the capacity in 

 the circuit influenced the breaking down effect in air and in oil. Without 

 external capacity boiled linseed oil broke down between small spheres — 

 at a distance of one millimeter — under a difference of potential of 

 20,000 volts. A sheet of microscopic cover glass ^^^ of an incli in 

 thickness was not, however, perforated, and a sheet of mica joVo ^^ ^" 

 inch in thickness also withstood this difference of potential. When a 

 capacity of 10,000 electrostatic units was employed no difference could 

 be observed in the breaking down effect. When therefore a powerful 

 storage battery is employed to generate a difference of potential, it 

 apparently serves the purpose of maintaining a full difference of potential 

 between terminals, and an external capacity does not influence the break- 

 ing down effect. 



The experiments I have described on the electric strength of thin layers 

 of glass showed tliat a number of Franklin plates one tenth of an inch in 

 thickness could be charged in multiple and discharged safely in series. 

 I had constructed at first an apparatus similar to that described by Plante. 

 I speedily found, however, that the design employed by him would not 

 enable me to utilize the full energy of 10,000 cells, for short circuiting 

 would occur in the apparatus. The form described by Plante was there- 

 fore replaced by the one shown in Plate I., and in plan and elevation in 

 Plate II. (Figs. 1 and 2). 



Figure 1, Plate II., represents an elevation of one end of the appara- 

 tus, and Figure 2 a plan of this end. The condensers are shown at (7 in 

 Figure 1, but are not re[)resented in Figure 2, in order that the con- 

 nections may be more clearly seen. In Figure 1, ^ is a handle which 

 moves by means of the link work D both levers L and L' at once. 

 When the lever L is lowered, conductors running along a and «i are 

 brought in contact with the coatings J^ and jE' of the plate condensers, 

 C connecting them in multiple. At the same time the lever L' is lifted. 

 When, however, the lever arm L is lifted, L' descends until the diagonal 



