According to this view, the apparent maximum of diatoms in late 

 September was a minor peak due to the earlier appearance of the 

 maximum of one fomiJ Stephanodiscus, 



A natural corollary of this conclusion regarding the abundance 

 of diatoms after the close of the season in 1930 is that the autumn 

 maximum was greater than that of spring. Because of the probable inac- 

 curacy of the figure for average abundance in late September, it is 

 unsafe to reach that conclusion on the basis of the data as they stand. 

 However, the circumstantial evidence cited in the preceding paragraph, 

 and the undoubted superiority in numbers in autumn over spring of 1929, 

 point toward the conclusion. 



Although diatoms may be regularly more abundant in autumn than 

 in spring in Western Lake Erie, the condition is not common to all shallow 

 lakes. As early as I89U, Whipple reported on the seasonal distribution 

 of diatoms in lakes and reservoirs of Massachusetts, Comparing deep and 

 shallow lakes, he found that the deep lakes showed two distinct periods 

 of abundance, one in spring and another in late autumn or winter, while 

 the shallow lakes showed little or no production in autumn. Certainly 

 this is not the case in Western Lake Erie. Whether or rot the autumn 

 maximum regularly exceeds the spring niaximum, the data for 1929 and 1930 

 show two periods of abundance very definitely. 



Whipple believed that the two periods of abundance of diatoms 

 in the deeper lakes were associated with the two periods of circulation 

 of . the water; that nutritive materials for the plants became isolated in 

 the stagnant lower water, and when they became available again, diatom 

 production was increased. While this interpretation fits in well with his 

 data, it can not explain autumnal increases in shallow lakes where there 

 is no stagnation in surmer, Tressler and Domogalla (1931) called atten- 

 tion to this fact in connection with their study of Lake Wingra, 

 Wisconsin, In this shallow lake (Maximum depth, U.25 meters), the diatoms 

 had a definite bimodal distribution, although the abundance in spring was 

 greater than in autumn. Marsh (1903, p. lU) believed that Whipple's 

 theory explained the seasonal distribution of diatoms (particularly 

 Cyclotella) in Lake Winnebago, but this co\ild not be the case, because 

 the lake lacks thermal stratification in summer (Marsh, page 6), 



A supplementary theory of diatom periodicity concerns the 

 effect of flood and drought on the concentration of nutritive salts. 

 Pearscill (1923 and 1932) is the principal advocate of this theory, although 

 earlier workers (some of whom were not quoted by Pearsall) laid the foun- 

 dation for it. Transeau (I916) reported that, contrary to the accepted 

 view, the salts are most concentrated in siu-face waters (of Illinois) in 

 spring and autumn, when the levels are highest, and that this is also the 

 time of most abundant fruiting of algae. Hodgetts (1921-1922) found a 

 varying relationship between water level and the amount of dissolved 



165 



