902 BUIvLETIN OF THE BUREAU OF FISHERIES. 



Since the determination of May 3 on untreated water agrees approximately 

 with the figures obtained several days earlier, it may be assumed that the water 

 does not vary beyond the limits shown in table i. This variation, however, is 

 considerable and as two samples could not be determined at the same time, it is 

 not known exactly what the condition of any treated water sample was at the 

 time it entered the troughs. It is known approximately, however, and it is 

 apparent that the 2-liter flow water is considerably but not completely corrected 

 for nitrogen and that much oxygen has been added but not enough to air- 

 saturate with oxygen. At this point the pump ceased to deliver water on ac- 

 count of clogging with sand. No more determinations were made until May 

 14. In the meantime a small hydraulic pump was connected with the well 

 and succeeded in pumping some 4 gallons per minute. This pump, however, 

 changed the aeration of the water considerably. 



The electric pump had its pumping cylinder entirely immersed and some 

 30 feet below the surface of water, so that all pipes were filled with water under 

 pressure instead of suction. There was therefore no opportunity for atmos- 

 pheric air to enter the pipes. The hydraulic pump on the other hand was 

 located on the surface of the ground and had a suction pipe some 22 feet long. 

 Though no leak was discovered, the pump delivered gas in large bubbles, much 

 more in quantity than ever came from the electric pump. This gas, or air, 

 must have entered the suction area at some point and though insufficient to 

 stop the pump, modified the air content in an interesting way, as shown by 

 comparing the "untreated" samples in table iii with table i. Oxygen has 

 been increased, while the nitrogen has not been materially changed or has even 

 been reduced somewhat. The explanation is found in the atmospheric air which 

 gains access to the hydraulic pump. The water having scarcely any oxygen 

 loses little or none in the suction pipe, but takes up considerable in the pressure 

 pipes between the pump and the point of delivery, on account of air taken in 

 at the suction and propelled in company with the water past the pump, where 

 it is then under pressure. The water having an excess of nitrogen must lose 

 considerable in the suction pipe on account of the reduction of pressure. This 

 nitrogen which comes out of solution remains within the pipe as free bubbles 

 and together with the atmospheric air sucked in, passes on with the water past 

 the pump when the pressure then causes nitrogen to be forced back into solution 

 in the water. The resultant of these two opposite processes is evidently a 

 slight diminution in the nitrogen content. This is reasonable, since the suction 

 below the pump is greater than the pressure above it. 



It is thus seen that although a leaky suction pipe in pumping systems 

 usually injures the water from a fish-cultural point of view, in this case it 

 improved it somewhat, by adding oxygen and subtracting nitrogen. It did so 

 because of the great length of suction pipe, the small head pumped against, and 



