Canon Moseley on the steady Flow of a Liquid, 195 



To determine approximately the value of 7, I have taken the 

 group of experiments 160. I have no reason to believe that these 

 were made with greater accuracy than any others. The value 

 of y which I have obtained from them is indeed certainly too 

 great. I have obtained it as follows : — 



By those experiments 



r 3 =-0643, w =l-303, i; 8 =M28; 

 •\ by equation (13) 



1-128 = l-303e-° 643 v; 



/1343\ 



But 



h 4nS)"- uimis > 



•14422315 00/lor oor , 

 , 7 = — — — — =2-2435=2-25 nearly : 

 •0643 J ' 



.-2-25r 



(W) 



The values of v } , v 2 , v 3 , as calculated by this formula for 

 eighty-eight experiments, are given in Table I., and also the cor- 

 responding values as given by experiment. The accordance is 

 approximative only, but it is undeniable. It is least apparent 

 in the last group of experiments, where the diameter of the pipe 

 was the largest (-J a metre), and where the head of water was small 

 and the velocity low. In those experiments the errors of obser- 

 vation were likely to be the greatest. In all, the difference be- 

 tween theory and experiment is greater at distances near the axis 

 of the pipe than further off. I do not know whether the chance 

 of errors of experiment is greater there than elsewhere. Expe- 

 riments on the efflux from a pipe are in their nature more reli- 

 able than any on the velocities of the liquid at different distances 

 from its axis can be. 



To determine approximately the efflux in terms of the velocity 

 v of the central filament, let Q K represent the volume of liquid 

 which flows per unit of time from a pipe whose radius is IX and 

 length I under the head of water h, 



Q R = I (2irrdr)v = 27r 1 vrdr ; 



•JQ Jo 



or, by equation (12), 



Q R = 27rv l e~ h- rdr. (15) 



02 



