30 THE LIFE OF DEVILS LAKE 



several inches thick occurs at the bottom, it is usually possible to 

 definitely establish this fact, but if this stratum is very thin, or if 

 oxygen is present at the bottom, but in quantities much smaller 

 than at higher levels, it is practically impossible to obtain duplicate 

 readings checking each other within less than 25 or 30%, Espe- 

 cially is this true in wind.y weather, when the motion of the boat 

 increases the currents generated in raising and lowering the appa- 

 ratus. 



There is some question regarding the accuracy of the Winkler 

 method, which was emploj'ed in this study, for the determination 

 of dissolved oxygen. Birge and Juday (1911) found a fairly close 

 agreement in the results of the Winkler and boiling out methods. 

 They were working, however, on waters, none of which approached 

 Devils Lake in salt concentration. In waters of such high concen- 

 tration as the latter, the iodiraetric method employed by Winkler 

 is liable to be affected by the various salts present. Relative 

 thereto, Dr. Heath, chemist at the station in 1914-15, says in ms. 

 notes on his results : 



"For a variety of reasons this method appears to be inaccurate 

 when applied to such a salt solution as Devils Lake water . . . 

 The presence of nitrates in the water causes the Winkler method 

 to give too high results for oxygen, and nitrates may interfere by 

 a variety of chemical changes. When the amount of nitrate is 

 known a correction may be applied, but even then there are so 

 many complicating factors that the method is unsatisfactory. It is 

 probable that the other substances present in the water also inter- 

 fere with this determination to a certain extent." 



The results obtained by the Winkler method, however, agree 

 well with those obtained by other workers elsewhere and with 

 what might be expected in Devils Lake. The boiling method, more- 

 over, is tedious, and difficult to use in the field, for which reasons 

 the Winkler method has been used and probably gives results which 

 are sufficiently accurate for comparative purposes. 



During the winter, when the ice sheet reaches an average 

 thickness of one metre, there is little opportunity for aeration, and 

 by spring the oxygen in the lower levels is greatly diminished if 

 not entirely absent. With the melting of the ice and influx of a 

 considerable body of fresh water from air and melting snow, the 

 oxygen content rises sharply, gradually diminishing again during 

 the warm days of summer. After a few hot still days, which but 

 rarely occur, the bottom water may be almost or entirely free from 

 oxygen, but a fresh breeze quickly changes this condition and restores 

 oxygen to the lower layers. With the advent of cool fall weather 

 the oxygen content again rises, remaining at a maximum until the 

 lake is once more sealed by ice. 



