DoWLiNG — Steady and Turbulent Motion in Gases. 



393 



representing a velocity 16, and the axis of diameter at a point '6 cm. to 

 the left. Thus the results may be expressed by either of the equations — 



V, = 26-6 {d+-&) . . . (7) 



or Vc' = {V- 16) = 26-6rf. . . (8) 



Three possible explanations suggested themselves. First, that it was 

 due to some irregularity introduced by unevenness on tlie tube-walls. That 

 this was improbable seems indicated from the fact tliat the effect was best 

 marked on the wider tubes, and was very indistinct on the narrower ones 



Y\% 3. 



The electrical quantities could be measured with confidence even at velocities 

 of 40 on the I'l em. tubes ; and their variations were certainly very 

 small where the discontinuity ' kink ' occurs (Curve II). Thus the unevenness 

 of a tube can hardly be the cause of the discontinuity. 



The second possibility is that it is due to some persistent eddies set 

 up at the mouth of the tube when the air enters and continuing inside. 

 In that case the effect would depend on the length of the experimental tube. 

 Varying lengths were used, no attempt at uniformity in this direction being 

 aimed at, yet the equations given above (7, 8) hold for all. 



8n 2 



