The dates on which one or both of the determina.tions of free carbon 

 dioxide appear to be erroneous are as follows: July 10 at Station l8: June 21 

 and July 9 at Station 59A; June 11 and 2$, and Aigust 9 at Statirn 76. The 

 questionable samples show too great a deficiency or too small an excess of car- 

 bon dioxide for the corresponding pH readings and oxygen content. If the 

 remaining sa;nples are accurate, they show that the water was commonly deficient 

 in free carbon dioxide. The pH values of the surface water tend to confirm 

 this finding; "Wily two of the 13 determinations were below pH 8.0, and the 

 lowest was pH 7.9. Eight were pH 8.3 or more. Only seven of the bottom read- 

 ings were 8.0 or above, and two were as low as 7.6. The low readings (7.6-7.7) 

 were correlated with partial depletion of ®xygpn in the lower water. Methyl 

 orange alkalinity (in terms of calcium carbonate) ranged from 86 to 95 parts 

 per million: 17 of the 26 samples were above 90. 



Certain general conclusions might be drawn from a study of the data, but 

 it seems advisable to present the more complete data of 1929 and 1930 before 

 attempting to draw the conclusions. 



Season of 1929 



In 1929 chemical s amnles were taken at eight stations in the Island Sec- 

 tion as follows: I8, 37A," 59A, 82, l58, 68, 75, and 8F. The location of 

 these stations may be seen in Fig. 1. Most complete data were obtained at 

 Stations 37A, l58, and 8f, and since conditions were found to be qjilte uniform 

 over the entire area, only these stations will be considered in detail. 



Station 37A. Samples were taken at this station on nine dates in the 

 period from late May to raid-October. It is recognized, of course, that 

 samples were not taken frequently enough to detect all of the changes in 

 chemical conditions which took place during the season, but it is believed 

 that the major trends are shown by the data. Samples were usually taken at 

 surface and bottom; on two occasions only surface samples were taken. The 

 data are given in Table l5 and Fig. 11 is a graphic representation of the 

 data from the surface. 



The mean depth at, this station was lh.2 meters. On every date for which 

 data are available, the top and bottom temperatures were very nearly -the same, 

 indicating that the water was frequently mixed from top to bottom. For that 

 reason only minor differences in chemical conditions between surface and bot- 

 tom were noted. Oxygen content of the surface water varied from 10.5 to 7.8 

 parts per million. The water was never completely saturated but was more than 

 90 per cent saturated on every date but one. Oxygen was most abundant at the 

 beginning and end of the season, when the temperature was low and solubility 

 of the gas high. The lowest point was reached in early July, but since this 

 was not the time of highest tehperature of the water, some factor other than 

 reduced solubility must have been involved. This is clear from the deep notch 

 'in the curve of satirration, which reached the low point of 8[i per cent. The 

 most probable explanation is that the oxygen was being used rapidly in decompo- 



63 



