compared mathematically for comparable periods of time. In the follow- 

 ing discussion, emphasis will be placed on differences between the 

 curves, rather than on their similarities. 



Turning our attention to spring, it is obvious that the 

 curves disagree with respect to actual abundance and the times of 

 maximum abundance. The count at the time of greatest abundance in 

 1930 was nearly twice as great as the corresponding count in 1929. 

 The difference was due in large part to the great abundance of 

 Rhizosolenia in 1930, This form made up more than half of the total 

 count of diatoms in late May, whereas in 1929 it was extremely rare. 

 In spite of the great abundance of Rhizosolenia in 1930, it was not 

 alone responsible for the early appearance of the maximum of the 

 diatoms. This statement may be verified by inspection of Table 5ii, 

 Even if Rhizosolenia were disregarded, the diatoms viould have reached 

 their maximum in late May rather than in early June, as in 1929j but the 

 count would have been almost the same as in early June, 1929. This 

 shifting of the time of greatest abundance possibly was due to the 

 higher water temperatures in 1930 as compared to 1929 (Figs, 8 and 9). 

 It is generally recognized that the maximal production of diatoms 

 occurs at times of rather low temperatures (Steuer, 1910, p, 538), 

 hence we should expect that the spring maximum of diatoms would come 

 earlier in a warm season than in a cool one. However, Pearsall (1923) 

 and others have expressed the opinion that temperature in it'' elf is 

 of little importance in determining periodicity of diatoms. 



The investigation was not carried on for a sufficient length 

 of time in autumn to cover all of the autumnal period of abundance and 

 subsequent decline. In 1929» sampling was continued into late October, 

 but at that time the diatoms as a group showed no indication of declin- 

 ing numbers. In 1930, no samples were taken after the first few days 

 of October. From the curve in Fig. l5 it might be concluded that the 

 diatoms had reached their maximum in late September and were declining. 

 But the high point in the curve was almost entirely due to 

 Stephanodiscus, and there is some question regarding the accuracy of 

 that point. There appeared to be a local aggregation of Stephanodiscus 

 at Station 59A on September l8, and the true mean number of diatoms 

 for the Island Section was probably lower than the figure obtained 

 (2ii2 thousand units). Furthermore, reference to Table 5h shows that 

 the diatoms other than Stephanodiscus had rather low counts when the 

 season closed. In view of the fact that water temperatures in early 

 autumn, 1930 (Figs. 8 and 9), were in excess of those for the same time 

 in 1929, it seems probable that the autumn maximum of diatoms came 

 later in 1930 than in 1929, and is not included in the records at hand, 



163 



