380 



Scientific Proceedings, Uoyal Dublin Societi/. 



of assuming that the diseontimiity is due in fact to turbulence setting in ; 

 but at the time the experiment was carried out, the writer was unaware of 

 Gibson's results, and it consequently appeared all the more advisable to 

 examine whether the discontinuity might not be due to some " end-effect," 

 the air being ionised before entering the tube at A. A further experiment 

 was accordingly carried out with the tube AB prolonged beyond A, by a 

 stiff paper tube, tightly fitting in, and of very nearly the same diameter. 

 The air in this case was not ionised until it had settled, more or less, into 

 its state of flow down the tube. The fact that the discontinuity was again 

 obtained at the same velocity was taken to indicate that it resulted from the 

 supposed cause. 



Fig. i. 



Using this form of apparatus, it was not possible to work over a large 

 range of velocities, because of the limitations imposed by the gas meter. This 

 was capable of passing a maximum of only about two cubic feet per minute, 

 and was found to be unable to register at all at velocities much lower than 

 one-eighth of that amount. However, as this experiment had shown the 

 possibilities of the method, a new arrangement of apparatus was adopted. 



