146 UNDER-REGISTRATION OF METERS 



This is based upon the idea of measuring small quantities 

 of water in a graduate or glass measuring vessel, by the ordinary 

 laboratory method, aided by the discharge through standard 

 orifices as a secondary method to permit more rapid comparison, 

 and upon counting the revolutions of the primary moving part 

 in the meter for a short period to get the rate of flow that is 

 indicated by the meter. In this way indications of performance 

 can be obtained almost at once, and running a meter for a long 

 interval to permit sufficient water to pass to be accurately 

 weighed on the large scales ordinarily used in testing and to be 

 registered accurately on the dial of the meter will be unnecessary. 



The water to be used for test will be drawn from the ordinary 

 pipe system. Directly beyond the shut-off gate would be a con- 

 nection to a pipe running up to a spout or to a sink on a floor 

 above, whose sole function is to reduce and steady the pressure 

 and to eliminate the effects of water-ram. With very steady 

 pipe pressure this may be unnecessary. Ordinarily a little 

 water will escape through this pipe and the pressure beyond this 

 point will be steady, and will only correspond to the elevation 

 of the outlet. 



The water should then be taken through a simple pressure 

 filter, which need be only a rough strainer, to stop any sediment 

 which otherwise might interfere with the subsequent operations. 

 ' This may be followed by a coil in a tank used to control the 

 temperature of the water used for experiment. Not much is 

 known about the effect of temperature upon the performance of 

 meters, but the viscosity of water at the freezing-point is twice 

 as great as it is at summer heat; there are probably unequal 

 temperature expansions in the rubber and the metal in the meters; 

 and it would be interesting, and perhaps profitable, to try out 

 the performance throughout the temperature range of practical 

 work. For this purpose the coil may be packed with ice, or sur- 

 rounded with water into which a little steam is drawn to pro- 

 duce the desired tefnperature. A pressure gauge at this point 

 is desirable. 



The water would next go through a controlling gate. This 

 should be a simple and substantial mechanism, so that by turn- 



