396 Scientific Proceedings, Royal Dublin Society. 



the same law both when it precedes aud when it follows the Reynolds' critical 

 point. 



A few other experiments were carried out to obtain further information 

 on the origin of the second critical velocity. The first was made with a view 

 to trying whether it might be possible to obtain the ' discontinuity ' by 

 producing irregularity inside the long copper tube. As we have seen, this 

 tube yielded no sigu of the effect. A short piece of glass tubing fitting 

 closely was pushed down close to where the air was ionised and just to the 

 left of the point A. This exaggerated the effect of a paper tube in the case 

 of the glass tubes. No break appeared on the curve, however, at any point 

 near the velocity expected. This negative result seemed to dismiss almost 

 completely any explanation based on irregularities in the bore of the tube. 

 It was accordingly sought to investigate the phenomenon by an altogetlier 

 new metliod. 



Part VI. 



Experiments on tlie Skin Friction due to Air flowing in Tubes. 

 Let us consider equation (9), above. It may be written 



») Ur =—«•,»■.., (9a) 



\(jrjr=a a 



wliere the left-hand term represents the tangential force (per square centi- 

 metre) on the ttibe due to the moving gas. 



Let us calculate the possible magnitude of this force for velocities such as 

 we have been using. 



Consider a tube 60 centimetres long, 1'9 cm. diameter; and imagine a 

 blast of 50 cms./sec. average velocity to be traversing it. Suppose the 

 temperature is 17°C., the viscosity will then be -000179; hence the total force 

 on the tube in the direction of its length will be 



277 rt • ^ • u ^ = Stt., • r,n ■ I . ■ ■ (10) 



or 



or 8 X 3-14 X -000179 x 50 x 60, or 13-4 dynes, 



equivalent to 13-7 milligrammes weight. . . (F) 



This seemed a sufficiently large quantity to measure ; and an experiment was 

 devised whereby this ' skin-friction,' as the force may be called, was utilized 

 as the effect whose variations should indicate the various changes in the 

 flow of the gas. 



Fig. 4 represents the arrangement adopted. One scale-pan of a 

 Sartorius balance was removed, aud also the pan-arrester. This left a small 



