66 



intervals of several weeks, and that such pulses can also be traced 

 back into 1897 as far as the collections were made at weekly inter- 

 vals that is to the early part of July. A similar periodicity on the 

 part of the Mastigophora the greater part of which are also 

 chlorophyll-bearing is even more evident. Not only is this 

 periodicity present in this group, but it coincides approximately in 

 the location of its maxima and in their relative development with 

 that found in the Chlorophycece and BacillariacecB. The following 

 table, which gives the dates of culmination of the pulses of these three 

 groups from July 1, 1897, to April 1, 1899, will serve to demonstrate 

 this point more clearly, and a graphic presentation of the data will 

 be found in Plates I. and II. 



There are twenty-two of these recurrent pulses in the period 

 from July, 1897, to March, 1899. Of the sixty-six possible maxima 

 only five are missing, or at least not apparent in our data, and but 

 ten culminate on other dates than the one (of collection) most to be 

 expected. These ten in every case culminate either a week prior or 

 subsequent to that in which the other two groups reach their max- 

 ima. These divergences may be due to the error incident to the 

 interval of collection, and their approximation in time is still cor- 

 roborative of the tendency towards recurrent periods of growth. 

 These exceptions are no greater than might be expected to occur in 

 the unstable fluviatile environment and within the large margin of 

 error of the plankton method. 



There are twenty-one intervals between July 14, 1897, and 

 March 14, 1899, with a range in length of 20 to 42 days and an aver- 

 age of 28.95. The intervals in days with the numbers of instances 

 of each are as follows: 20 (1), 21 (3), 22 (1), 23 (1), 26 (1), 

 27 (1), 28 (7), 35 (3), and 42 (3), days. The effect of the weekly 

 interval of collection is seen in the preponderances at 21, 28, 35, and 

 perhaps at 42, days. There is evidently a tendency towards the 

 interval of 28 days. Nine of the 21 pulses are grouped about this 

 interval; 6, about that of 21 ; while 3 are at 35 and 3 at 42. If there 

 be such a tendency it is but natural that with a weekly interval of 

 collection there should also appear minor preponderances at 2 1 and 

 35 days. Traces of a similar rhythm may be found in the period of 

 weekly collections in 1896 (Pt. I., Table III.). 



In some instances the environmental conditions at these times 

 of departure are such as to suggest that they may have produced the 



