35 
is at a lower level than the summer one. In 1894 the interval of 
collection is too great to follow the seasonal distribution, but there 
are hints of summer and autumnal pulses. In 1896 there were no 
May collections, and the largest number, 6,060,665, appears June 
- 19, five minor pulses —on July 18, August 21, September 12, October 
11, and November 5—intervening before the hiemal pulse of 3,574,- 
028 appears on November 27. Other pulses follow on December 18, 
January 6, February 4, March 4, and March 17, before the vernal 
pulse of 1896 culminates at 105,440,858 on April 24. This is fol- 
lowed by minor pulses on May 18, June 11, July 18, August 8, and 
September 16, and by the hiemal pulse of December 3 of 346,982,- 
928*. The vernal pulse of 1897 appears April 27 at 6,207,473,520, 
but is surpassed by a pulse on July 14—principally of Melosira 
spinosa—of 11,459,289,600, and minor pulses then follow on August 
ee seprembper 29, October 26, and December 7 and 21. ‘The hie- 
mal pulse of this year is insignificant. In 1898 three minor pulses 
appear, January 21, February 15, and March 22, and the vernal 
pulse culminates May 10 at 3,865,257,360. Minor pulses follow on 
June 14, July 19, August 9, August 30, September 27, October 25, 
and November 22, and the hiemal pulse culminates December 15 at 
436,535,790, followed in 1899 by minor ones on January 10, Febru- 
ary 14, and March 14. 
Some of the pulses here indicated are due to the development of 
single species, as that of Melostra on July 14, 1897. Most of them, 
however, are composite, including a number of species. This is 
especially true of the vernal pulse, which in 1898 is due to the com- 
bined increase in Fragilaria virescens and F. crotonensis, Cyclotella, 
Astertonella, Navicula spp., and Synedra acus. Astertonella culmi- 
nates early in the vernal pulse and the majority of the others 
towards its close. Melosira varians is among these, but M. 
spinosa contributes less to this pulse than it does to later ones. 
Minor pulses are also composite, as, for example, that of August 9, 
1898, which is due to Melosira spinosa, Cyclotella, and Navicula. 
FACTORS CONTROLLING DIATOM PRODUCTION. 
The fact that many of these pulses represent the combined 
fluctuations of a number of species leads us to look for some factor 

* Filter-paper collections included in this and in following years. 
(4) 
