Mass of the Ions in the Electric Wind in Air. 



413 



and water before use, and the air was passed through strong 

 sulphuric acid and over phosphorus pentoxide before it entered 

 the apparatus. 



In the following table are collected all the reliable values 

 of V obtained. 



Micro- 



V4- 



V- 



i 



Micro- 



| 

 V+ 



V- 





anips. 









amps. 

 32 









32 

 1-6 

 32 



401 

 396 

 414 



407 



516 

 506 

 545 



560 



Cune. 

 i> 



f Double 

 \ position. 



430 



427 

 421 

 415 



547 



547 

 547 

 514 



f Double 

 \ Position. 



» 



420 

 412 

 401 

 430 

 417 

 404 

 433 

 436 

 419 



513 

 524 

 547 

 526 

 515 

 510 

 602 

 562 

 557 



>> 



3 : 2 1 

 1-6 ] 

 3-2 } 

 1-6 ] 

 3:2 \ 

 1-6] 



401 

 396 

 423 

 390 

 376 

 415 

 390 

 407 

 422 



534 

 524 

 520 

 629 

 620 

 540 

 551 

 539 

 541 



i) 



Those marked " curve " are deduced from complete sets of 

 readings such as those from which curves IV. are plotted. 

 "Double position" means that the pressures were measured 

 at two values of z only, situated w T ell within the straight 

 portion of the curve (usually at 5*18 and 3" 63 centim. 

 respectively). It will be seen that most of the determinations 

 were by the latter method ; this was for the following reason. 

 In taking readings for a complete pressure curve I usually 

 went up and down the range of z several times so as to get 

 several determinations in succession of what ought to have 

 been the same curve. These seldom coincided, however, 

 when plotted ; yet their disagreement was not haphazard, as 

 each separate curve was usually fairly smooth, often with a 

 gentle bend up or down indiscriminately. This could be 

 accounted for by supposing that a small part of the pressure 

 was variable and independent of the distance from point to 

 ring ; i. e. of that part of the pressure from which V is 

 calculated. Such a variation of pressure, if it occurred 

 gradually during the taking of the readings for a long curve, 

 would obviously alter the slope of the curve, though there 

 would be no corresponding alteration in V. It was there- 

 fore important to take the readings for large and small 

 values of z alternately, and as quickly as possible ; hence 

 the " double position " method. In illustration I give the 

 details of a complete experiment. 



