M 



364 Messrs. J. D. Fry and- A. M. Tyndall o?i 



4. Lastly, the presence o£ x might be the result of a slight 

 suction at the static pressure tube. It is well known ttait if 

 a piece of thin- walled tubing is placed in a stream so as to lie 

 with its long axis perpendicular to the flow, such a suction 

 does occur, and it may even amount to as much negative 

 pressure as a pitot placed at the same point would give of 

 positive pressure. It has, however, generally been supposed 

 that this is removed by either providing the tube with a wide 

 flange at its mouth or by placing it, as in the present work, 

 with its end flush vvitn the inner surface of the pipe. In 

 support of this there is the work of Heenan and Gilbert 

 (Proc. Inst. Civil Eng. vol. cxxiii.) and Threlfall (loc. cit.), 

 and also the fact that in the present work K is approximately 

 the same as K x within the range of velocity used by most 

 experimenters. 



It is conceivable, however, that suction is not completely 

 eliminated in these ways. If, for instance, a residual suction 

 is present which is a function of a power of the velocity less 

 than the square, it might easily only be comparable with the 

 pitot pressures at very low velocities, and might therefore 

 have escaped previous detection. 



For instance, making the very arbitrary assumption that.?; 

 is proportional to v at all velocities — the values of K 

 calculated by correcting all the pitot pressures accordingly 

 does not differ more than 1 per cent, from TOO throughout the 

 whole range from 6 to 2000 cms. per second. 



On the other hand, if there were suction at the static 

 pressure tube, a change in the size and shape of that 

 tube lni^ht be expected to affect the value of ~x. But no 

 difference of pressure was observed when the gauge was 

 connected to any two of the following holes made at the 

 same level in the side of the tube — circular holes 1 mm. and 

 4 mm. diameter, rectangular holes 1*5x3 and 3x1*5 mm. 

 respectively. On the whole, therefore, the suction hypothesis 

 does not seem feasible, though the fact that 1c \< proportional 

 to the rate of flow drives one to locate the effect at the walls 

 of the pipe and not at the pitot itself. It is possible that 

 experiments on the pitot pressures in pipes of different sizes 

 may throw some light on the subject. 



Pitots of different shapes. 



Before it was realized that the pressure readings across a 

 diameter were not in accordance with theory, an attempt was 

 made to examine the motion of the air at the side of the pipe. 

 Even when eddy motion is present the flow close to the 



