MEASUREMENT OF DISCHARGE 221 



fairly large pipes an error not exceeding 1*5 per cent, is to be expected. 

 With small pipes, one of the other types of meter described in Art. 196 

 may be used. 



(5) A chemical method of determining the discharge is described by 

 F. Van Iterson. 1 A given weight per minute of sodium thiosulphate 

 being introduced into the pipe, a sample of the discharge is taken at 

 some point nearer the exit, and the quantity of chemical present is 

 measured. The chemical being diluted in proportion to the volume of 

 discharge, a measure of the latter is obtained. The sample is titrated 

 with standard iodine, starch being used as an indicator. Several cocks 

 arranged at different points on the circumference of the pipe should be 

 used for withdrawing the samples, so as to obtain as far as possible a 

 fair sample of the mixture. This method has been used with success for 

 determining the discharge of sewage pumps at La Haye. 



Where a colourless discharge is expected, coloured liquid may be 

 introduced into the pipe in known quantities, and the colour of the 

 discharge compared with that of standard admixtures. The method 

 is, however, only tc be looked upon as giving approximate results. 



ART. 69. RELATION OP DIAMETER OF PIPE TO QUANTITY 

 DISCHARGED. 



jn 



With rough pipes, since Q = 7T ~- . v, and since the loss of head H in 

 any length I is given by 



H^ii/r 



201 mj 



. J*gHm 



fl 



~. d> where C" = ,. 



I z/ 4 



For smooth pipes, the loss of head is proportional to a lower power of 

 the velocity than the second, and to a power of the diameter slightly 

 different from the first, i.e., 



v n 7 



Le Gtnie Ciml t Paris, vol. 44, p 411. 



