440 PROCEEDINGS OF THE AMERICAN ACADEMY. 



the case of his apparatus much less than in induction apparatus. This 

 remark applies with great force to the modern use of transformers for 

 the production of high tension effects. With the improved form of 

 rheostatic machine which I have had constructed, one third of a horse- 

 power will produce the effects which have hitherto required the em- 

 ployment of thirty to forty horse-power. 



Up to the point of one million and a half volts the length of the elec- 

 trical discharge in air appears to be closely proportional to the electro- 

 motive force. When this voltage is exceeded, the length of the spark no 

 longer increases in proportion to this force. For instance, an electro- 

 motive force of approximately three million produces a spark of about 

 seven feet in length, two hundred and ten centimeters, when it should 

 excite one at least ten feet in length. The reason of this diminution 

 in length is readily seen when the operation of my apparatus is examined 

 in the dark. From both terminals and from the conductors to the ter- 

 minals there is a luminous brush discharge to the w-alls and floor of the 

 room. The main discharge between the terminals is so to speak shunted 

 through the air, which breaks down with facility at such high voltages. 

 The high electromotive force exerts a similar action to that of diminished 

 air pressure. 



It is thus interesting to compare these conditions further. I therefore 

 connected in multiple two Leyden jars, of about five thousand electrostatic 

 units each, with the poles of five thousand of my cells, and measured the 

 disruptive discharge in a receiver which contained air at ten centimeters 

 pressure. The length of the discharge at this pressure was approximately 

 eiglit centimeters, when the jars were charged with ten thousand cells. 

 The length of this discharge was fourteen centimeters. Here twice the 

 voltage did not produce twice the spark length. The diminution of length 

 was evidently due to a species of shunting through the air, which had 

 become a relatively good conductor. Indeed, one sees the luminous area 

 of discharge on the positive terminal extend farther and farther in the 

 case of pointed terminals from the point of the terminal. 



In ordinary atmospheric air, therefore, the same increase of electrical 

 conductivity takes place under the action of great electromotive force. 

 It is difficult to make exact measurements of the action of such powerful 

 discharges; but the following qualitative experiments, I think, will illus- 

 trate my meaning. When discliarges produced by one million volts and 

 over are excited between terminals six feet apart in tubes filled with 

 ordinary water, the tubes are speedily burst, and when the phenomenon 

 is carefully examined it is perceived that disruptive sparks occur on 



