TESTING AT LOW FLOWS 145 



filtered or ground water will give a record of smaller flows than 

 can be shown by meters now commonly used. 



The use of compound meters for large services and the im- 

 provements that are being rapidly made in the design of such 

 meters is rapidly bringing about a better condition with respect 

 to the larger consumers. But it must always be borne in mind 

 that the representative domestic consumer is one whose meter 

 records about 100 gallons per day. On the best available evi- 

 dence, the actual average quantity of water drawn is con- 

 siderably more than 100 gallons per day, possibly as much as 150 

 gallons per day. The average meter records that part of it 

 which flows at relatively high rates of draft, amounting in all 

 to about 100 gallons per day, and it fails to record the drop-by- 

 drop leaks from water-closets and fixtures, amounting to some 

 very substantial quantity, but varying from house to house and 

 also from system average to system average; so that whether 

 as a general average it amounts to 50 gallons per day, as the 

 Committee on Meter Rates suggested that it might, or to a larger 

 or a smaller quantity, cannot be certainly determined. 



Testing Meters at Low Rates of Flow 



It is obvious that testing meters, especially domestic meters, 

 at low rates of flow goes to the heart of the under-registration 

 problem. Testing meters at ordinary rates has been brought to a 

 state of perfection where little can be suggested to improve it, 

 but testing meters at low rates has been neglected. 



It has been neglected perhaps primarily because of the 

 difficulty of making tests with the equipment that has been 

 available. It may be fairly supposed that if a rapid and easy 

 method of making tests of the performance of meters at low 

 flows were available, that information would soon be collected as 

 to the performance of different kinds of meters when new and 

 after certain periods of service, and this would certainly react 

 upon the manufacturers and tend to beneficial improvements. 



The following outline is suggested of a simple apparatus to 

 permit tests of this kind to be made with rapidity and all needed 

 accuracy. 



