the, Value of the Pilot Constant, 365 



walls must still be stream-line and must break up into eddies 

 at some unknown distance from the walls. It seemed that 

 the place where this change occurred and its relation to the 

 mean velocity of the air, could be determined by using a pitot 

 which was narrow enough in section to approach the walls 

 very closely. The above results show this is not possible 

 unless the pressure readings are corrected for the x effect, 

 which in practice is impossible to do very accurately ; but 

 at the same time, the attempt brought out a new and 

 interesting fact. 



The results of previous experimenters with various pitot 

 tubes have led to the generalization that the size and 

 shape of a pitot has no effect upon the value of the 

 pitot constant K. The authors found, however, that a very 

 fine pitot does not show the same pitot pressures as a larger 

 one say 1 mm. in diameter. For instance, values of the 

 pitot constant at various velocities have b^en taken with a 

 very fine rectangular pitot 0*2 mm. by 2*0 mm., and at high 

 velocities this gives a mean value for K of about 1*10. This 

 pitot was obtained by flattening the end of a brass tube and 

 grinding down its edges to a thickness of 0*05 mm. 



A pitot with about the same opening 0*1 mm. by 2'0 mm. 

 but thicker walls each 0*2 mm. gave, however, a normal 

 value K = 1'00 at high velocities. 



A circular glass pitot '111 mm. diameter and with walls 

 *027 mm. in thickness, also gave a value of K several percent 

 above unity. 



Now close to the edges of the pitot, the pressure due to the 

 current of air will fall off rapidly and will be considerably 

 less than that at its centre. This edge effect has been 

 thoroughly investigated by Stanton (loc. cit.) for the case of 

 plates placed in uniform currents of air. In an ordinary 

 pitot it has no effect on the observed pressure, which is the 

 average pressure over the internal area of the end of the 

 pitot tube, because either the walls of the pitot are relatively 

 thick or its area is considerable. In the above pitots with 

 high constants, the area influenced by this edge effect was 

 evidently a measurable fraction of the whole area and the 

 resultant pitot pressures were, therefore, all too small. 



In support of this view the authors found that if a small 

 piece of mica was fitted on flush with the end of one of these 

 pitots so that the pitot was then a circular plate 2 mm. diameter 

 with a slot at its centre, the value of K obtained was normal 

 1-00. 



Further, when the wedge pitot was touching the side of 

 the pipe, the pitot pressure was actually greater than when it 



