512 REPORT— 1902. 



6. Further E.tperiments hy the Thermal Method on the Variation of the 

 Critical Velocity of Water tvith Tenqierature. By Howard T. Barnes, 

 M.A.Sc, D.Sc. 



In a paper communicated by the author with Dr. E. G. Coker to the British 

 Association at Glasgow it was shown from a series of experiments by a new 

 thermal method on the upper limit of stream-line flow that this point, at which 

 stream-line motion breaks down, apparently does not follow the temperature func- 

 tion of Poiseuille exactly. A second series of measurements by the same Inethod 

 for a diflerent pipe has given identically the same result. 



Realising how contrary to theory the result of this experimental investigation 

 .seems to be, it was decided both by Dr. Coker and the author to carry out separate 

 determinations involving measurements of the hnuer limit of stream-line flow and 

 the dependence of the same on the temperature. Since the theoretical equations 

 apply to this lower limit, which is the point where the law of resistance changes 

 and the true criterion is obtained, a rigid adherence to the law of Poiseuille neces- 

 sarily follows.' 



In place of starting with water free from initial di.sturbance, as in our previous 

 experiments, and measuring the velocity at which stream-line motion breaks down, 

 water in a disturbed state was allowed to flow through a pipe of sutiicient length 

 to allow of the dying out of the eddies for velocities less than the critical. By 

 surrounding the end of the pipe with a water-jacket at a slightly difi'erent tempera- 

 ture from that of the water flowing through, and using the thermometer as before, it 

 has been possible to determine : («) tlie velocity at which all the eddies die out 

 and stream-line motion is the stable flow ; {h) tlie velocity at which the motion is 

 entirely sinuous with no tendency to form stream-lines ; and (c) the intermediate 

 velocity where by the oscillation of the thread of the thermometer the formation 

 of stream-lines and the breaking up into eddies were seen to follow in regular 

 intervals. These experiments were made by the author partly to test the efticacy 

 of the thermal method for determining the lower limit of stream-line flow with a 

 view to extension to other liquids. 



The result of this and the previous work may be summarised briefly as 

 follows : — 



1. An experimental verification of the law of Poiseuille for the temperature 

 coefficient of the lower limit of stream-line flow. 



2. That there is a small divergence from the law of Poiseuille in the case of the 

 temperature coefficient of the upper limit of stream-line flow. 



3. The reformation of stream-lines, shown both by the thermal method and 

 colour-band tests, for velocities above that at which eddies first make their appear- 

 ance. 



4. The attainment of exceedingly high velocities of stream-line flow for certain 

 sizes of pipes under steady conditions. 



FRIDAY, SEPTEMBER 12. 



Department II. — Astronomy and Cosmical Physics. 



Chairman : Professor A. Schuster, Ph.D., F.R.S. 



The Chairman delivered the following Address: — 



Otjr proceedings to-day constitute an innovation and require a few words of ex- 

 planation. AVhen, a few years ago, some astronomers felt that our Association 

 bestowed au insufficient share of attention on their subject, an easy remedy sug- 

 gested itself in the formation of a special sub-section devoted to that subject. Such 

 a sub-section was accordingly organised at Bradford and Glasgow, but for reasons 



' Osborne Reynolds, Phil. Trans., 1883 and 1895. 



