244 ELEMENTARY LESSONS ON [CHAP. iv. 



length of spark between two conductors increases with 

 the difference between their potentials. It is also found 

 to increase when the pressure of the air is diminished. 

 Riess found the distance to increase in a proportion a 

 little exceeding that of the difference of potentials. Sir 

 W. Thomson measured by means of an " absolute elec- 

 trometer " (Art. 261) the difference of potential necessary 

 to produce a spark discharge between two parallel plates 

 at different distances. His precise experiments confirm 

 Riess's observation. Thus, to produce a spark at -I of 

 a millimetre distance, the difference of potential must be 

 2-7 (arbitrary) units ; at *5 millim. 7-3 units ; at I millim. 

 1 2 -6 units; and at 1*5 millims. 17*3 units. De la Rue 

 and Mtiller have found with their great battery (Art. 174) 

 that with a difference of potential of i ooo volts the strik- 

 ing distance of the spark was only -0205 centimetres (or 

 about T^ of an inch), and with a difference of 10,000 

 volts only '2863. Their 1 1,000 silver cells gave a spark 

 of "3378 centim. (about y of an inch) long. To produce 

 a spark one inch long, through air at the ordinary 

 pressure, would therefore require a difference of potential 

 exceeding that furnished by 70,000 Daniell's cells ! 



The length of the spark differs in different gases, being 

 nearly twice as long in hydrogen as in air at the same 

 density, and longer in air than in carbonic acid gas. 



In rarefied air the spark is longer. Snow Harris 

 stated that the length of spark was inversely proportional 

 to the pressure, but this law is not quite correct, being 

 approximately true only for pressures between that of 

 eleven inches of mercury and that of 30 inches (one 

 atmosphere). At lower pressures, as Gordon has lately 

 shown, a greater difference of potential must be used to 

 produce a spark than that which would accord with 

 Harris's law. From this it would appear that thin 

 layers of air oppose a proportionally greater resistance 

 to the piercing power of the spark than thick layers, and 

 possess greater dielectric strength. 



