146 WAVE IMPACT ON 



lines in Fig. 2. Water was allowed to flow from an open tank 

 through, the upper horizontal length I of the pipe, through the 

 valve V. This valve being closed suddenly the corresponding 

 hammer pressure was measured by the indicator at the lower 

 end of the coil. The results of these experiments were very 

 satisfactory, the pressures as calculated from the formula 

 p=4c8 v Ibs. per square inch, and as obtained from the indicator 

 diagrams with ordinates multiplied by the corresponding 

 multiplying factor, agreeing, for values of I between 20 inches 

 and 60 inches, within about 3 per cent. Evidently if Z=0, 

 corresponding to normal impact on a joint which is full of water 

 the value of the multiplying factor becomes infinitely great, 

 and the indicator could only show the effect of a hammer 

 pressure if this were also infinitely great. 



A further check was obtained by taking diagrams under 

 similar conditions with T ^ and -fa indicator springs. Since 

 n for the ^j spring is twice as great as for the T 1 spring, the 

 pressures indicated with the former spring for values of I in 



the neighbourhood of 25 inches (cos xi * for fa spring, 



=1 for fa spring) should, if the reasoning is valid, be 

 approximately twice those recorded with the weaker spring. 

 The results showed that the pressures recorded by the stronger 

 spring were higher in a ratio which varied from 1*75 to 2*2. 



Adopting the formula as giving substantially accurate 

 results, it becomes possible to deduce, from an indicator 

 diagram, the true pressures in the pipe when a jet is suddenly 

 diverted into its open end, provided only that water-hammer 

 actually does take place. 



A preliminary series of experiments, carried out to settle 

 this point, showed that by slightly modifying the method of 

 application of the jet, two distinct types of diagram could be 

 obtained. The jet played instantaneously and normally on 

 to the open end, usually gave a diagram identical with those 

 obtained when the piston was in use (Fig. 6 a) indicating that 



