51 



which the species dwindles away and disappears in August to return 

 early in November. The winter pulse culminates December 14 at 

 8,159,250, at a break in the ice blockade. In 1898 the winter mini- 

 mum continues into April, and the vernal pulse appears May 10 at 

 253,960,000, rising with rocket-like suddenness from 390,000 of the 

 previous week, and declining the week following to 4,110,400. The 

 decline to the summer minimum is prolonged into July, and the 

 species does not reappear until October. The winter pulse begins 

 earlier than usual, on November 1, and is well sustained through 

 the month, culminating on the 29th at 2,254,000. The winter mini- 

 mum which follows, does not reach the low levels of that of summer. 

 This species has thus a characteristic distribution, the analysis 

 of which is by no means simple. The contrast between the summer 

 and winter minimum may be due to the low nitrates of the summer 

 and the larger amount in the winter (Pt. I., PI. XLIII.-XLV.), 

 which favor a proportionate development of this diatom, though 

 not every species shows this response. The two minima separate 

 the seasonal occurrences of this species into two periods of growth ; 

 a vernal, from March to June, and a hiemal, from October to Janu- 

 ary, the limits and relative development of each being somewhat 

 variable from year to year. The temperatures of the two periods 

 differ. Both are times of rapid change, of rise and fall respectively, 

 and the culminations of the periods of growth lie at widely sepa- 

 rated temperatures. The vernal pulses in 1896 and 1898 in which 

 years collections were frequent enough to locate them with some 

 degree of accuracy appear at 72 (April. 24) and 61 (May 10) 

 respectively, and in every year the vernal pulse appears during a 

 period of rapid change. The hiemal pulse, on the other hand, cul- 

 minates in each year after the winter minimum approaching 32 has 

 been reached, and in two years during the ice blockade. Tempera- 

 ture within these limits seems not to be a determining factor in the 

 pulses of this organism. The nitrates (Pt. I., PI. XLIII.-XLV.) 

 have been uniformly high (above 2 parts per million) whenever the 

 pulses occurred. In 1898 they decline abruptly (Pt. I., PI. XLV) 

 and remain at a low level throughout December, and in this month, 

 when usually Fragilaria attains its hiemal maximum, we find it 

 dropping to the unusual minimum of 20,000. The pulse which 

 began in November is cut off apparently by this unusual decline in 

 nitrates. Abundance in nitrates is not, however, in itself sufficient 



(5) 



