t)0AVLiNG — Steachj and Turbulent Motion in Gases. 377 



but the pressure-differences are of a very small onler, unless, indeed, 

 tubes of an abnormal length are employed. For example, Riedlev and 

 Grutermutli' experimented with pipes upwards of 10,000 to 50,000 feet long, 

 and about 1 foot in diameter. They had pressures of upwards of twenty 

 pounds to the square inch. Tliey, however, obtained only a few numbers — 

 mostly referring to velocities above the critical. 



Other experimenters used tubes down to one-eighth of an inch bore, 

 in which case the length could be cut down to between 30 and 100 feet, 

 still having pressures of some pounds per square incli. These experiments 

 were performed at different times, and all with a view to solving the problem 

 of the resistance to the flow of air through pipes. The experimenters were 

 not specially concerned with the problem we are now considering. 



To verify the applicability of Eeynolds' formula to the case of gaseous 

 flow, through tubes of widely different diameters, by methods involving 

 pressure measurements, would present great difficulties. Anotlier method 

 was therefore adopted. 



Part II. 



Principle of the method employed. 



It will be, perhaps, well to notice in what turbulent motion consists. 

 When a certain mean velocity (the " critical " velocity) is reached, the fluid 

 flowing in a pipe ceases to follow stream-lines. This condition of affairs 

 manifests itself in various ways. For example, the pressure-difference 

 required to maintain a certain velocity in a fluid above the critical point 

 is much greater in proportion to that required below that point. We may 

 notice that probably the turbulence sets in at the centre of the tube at tirst, 

 and gradually spreads itself over the whole cross-section. The net result 

 of the action is that a fluid particle on the average will follow a longer 

 path during turbulent motion in travelling between any two points of a 

 tube than in covering the same distance in " stream-line " conditions. For 

 our purpose this is the important point to keep in view. 



Now a gas that has been made electrically conducting by the in- 

 fluence of a radioactive substance will gradually lose its conducting 

 power, if removed from the vicinity of the ionising agent. This gradual 

 loss of conductivity is due to the ions in the gas becoming neutralised in 

 one or other of the following manners. (1) Ions of opposite charges may 

 recombine. (2j Ions will diffuse to the walls of the containing vessel : 



1 Gibson, loc. cit. 



8 L 2 



