from an Arc and from Incandescent Metals. 247 



t being the time required for the stream of gas to move through 

 a distance equal to the length of the terminal. 



We calculate p from the two determinations of the rate of leak 

 from B when G is to earth and when kept at a known potential, 

 the time t is easily found, so that we can deduce the value of v. 



A large number of determinations of the velocity of the carrier 

 was made in this way and the value found for the velocity was 

 quite constant as long as the nature of the arc remained the same. 

 When the arc was varied by increasing or diminishing the resist- 

 ance in series with the primary of the induction coil between the 

 secondary terminals of which the arc was formed, and therefore 

 the amount of current through the arc changed, it was found that 

 the velocity of the carrier also changed. When the amount of 

 current passing through the arc was increased the velocity of the 

 carrier was diminished and vice versa. The following numbers 

 refer to one determination of the velocity of the carrier. 



Eate of heat from B. 

 G to earth 50 scale divisions in 21" 



G at 20 volts 50 scale divisions in 52". 



From these numbers we find p 2 = *45. 



The velocity of the stream of air was 19 cms. per second. 

 Substituting in the above equation we get 



v = "015 cms. per second. 



(6) This determination refers to a case where the current 

 through the arc is relatively great. The velocity of the carrier 

 increases rapidly as the current through the arc is diminished. 

 The current through the arc was cut down step by step and the 

 velocity of the carrier determined for each state of the arc, and 

 values were found increasing from that given to "33 cms. per 

 second. It was difficult to maintain the arc with smaller currents 

 than that which gave the latter velocity. The velocity of the 

 carrier produced by the arc is small compared with that produced 

 by the Rontgen rays ; in this case Rutherford found for air a 

 velocity of 1*6 cms. per second. For the velocity of the carrier in 

 the conducting gas from a flame I found the value '2 cms. per 

 second, which rapidly diminished as the hot gas cooled down. In 

 conducting gas from an electrolytic cell Townsend found as low a 

 velocity as "0007 cms. per second. 



It is quite likely, however, that the initial carrier may be the 

 same in all cases, but with more uncharged masses travelling 

 with it when the velocity is smaller ; the rapid decrease of the 

 velocity in flame gas would point to this increase of mass as the 

 gas cools down. The variation of the velocity of the carrier in the 

 gas from an arc may be explained in a similar way ; when the 



