1 HE IRRIGA 7 ION' A GE. 203 



head than the centrifugal. Great wisdom should be used in selection 

 of rotary pumps to receive good wearing properties. 



In the large majority of wells the water gradually falls below 

 suction. In such cases the pit must be deepened and the pump low- 

 ered. One of the best devices for this consists of a "stirrup" to hold 

 the vertical centrifugal pump. This is thus freely suspended from 

 the top of the well and can be lowered while at work. In thus lower- 

 ing the shaft the best device I have noticed for preventing loss of 

 suction through too rapid pumping, consists of a pipe connected with 

 the delivery pipe just above the pump and turn-down into the well. 

 With such valve you can regulate the amount of water flowing back 

 into the well to keep the water up to suction. Such a device is often 

 useful to secure automatic action. Those who have suffered loss of 

 time from loss of suction will appreciate so simple a device. 



One of the common conditions is a group of wells each yielding a 

 small amount of water. These can be connected together if not too 

 far apart, but if there is a general fall in water level, making it 

 necessary to lower the pump to keep it within suction limit, it is an 

 expensive matter to lower these connections. It is not a very serious 

 matter if the wells are within a radius of thirty feet. Often the only 

 reason for a nest of wells is that the sand is so fine that it gives up its 

 water slowly. If you take sand and gravel, coarse enough to go 

 through a twenty mesh screen, water will pass through it so freely 

 that you can pump 100 miner's inches or more out of a seven inch 

 casing. In the same water area where the strata are unquestionably 

 all connected, you may yet strike sand so fine that much of it will go 

 through an eighty mesh screen. You take a quart can of this 

 material dry and you can pour into it almost as much water as into 

 the coarse gravel, but you can turn the glass over and not a drop will 

 run out. 



Where you have fine sand and yet are sure that the water is 

 present in large quantities, large wells should be put down and close 

 together and the group connected to one well, using a deep pit and 

 tunnels to the side walls. 



In many cases the wells are necessarily scattered a half mile 

 apart; no one will yield much water; and the general water level 

 lowers rapidly from month to month. Such an area is not a valuable 

 permanent supply, but may be worth very much in a dry year. 

 A central compressed air plant may work here with greater economy 

 than any other plant. Never use compressed air on a single well. 

 There is no question that it is highly efficient, but in such a case as the 

 above it would be most economical all thing considered. Where the 

 wells are scattered over a large area for the purpose of saving 

 conduits and to deliver good irrigating heads at convenient distances, 

 and where the area calls for constant pumping, I think the plan 



