273 



these exceptions no record exceeding 600 per m. 3 was made between 

 the dates named. Between April 20 and October 1 the minimum 

 records rarely fall below 600 per m. 3 , except in 1898, and the pulses 

 often culminate at 2,000-8,000. C. viridis var. insectus is thus a 

 planktont of the warmer season, and its seasonal distribution is 

 strikingly similar to that of the so-called var. brevispinosus. 



This form occurs in our plankton throughout the whole seasonal 

 range in temperatures, but only in small numbers and irregularly 

 below 60. Only 21 per cent, of the collections containing 

 insectus were made at temperatures below 60, and these contained 

 less than 3 per cent, of the total individuals. With the exceptions 

 of the pulses culminating at 43 November 23, 1897, at 1,200 per 

 m. 3 , and at 57 April 26, 1898, at 4,160 per m. 3 , no development of 

 this species exceeding 600 per m. 3 occurs below 60. All pulses of 

 more than 3,000 per m. 3 , excepting only the April pulse of 1898, 

 occur at temperatures above 70. The species reaches its greatest 

 development in channel waters during the period of maximum 

 temperatures, 70-80. 



The seasonal distribution of this form shows a few straggling 

 individuals in November-March during temperatures below 50, 

 and a meteoric rise to a vernal pulse in April-May as this tempera- 

 ture is passed and 60-70 arrives. This is followed by a series of 

 recurrent pulses, often of considerable amplitude, through Septem- 

 ber or until temperatures fall below 60, as in October, 1897. With 

 falling temperatures the drop in numbers to the winter minimum 

 is quickly accomplished. A comparison of the distribution in 1897 

 and in other years, shows a close correlation between the decline in 

 temperatures and the falling off in numbers of insectus. 



The relations which hydrographic conditions bear to the develop- 

 ment of insectus in channel plankton may be inferred from the 

 hydrographs on Plates IX.-XII, Part I., and from the data sum- 

 marized in the following table, 1894 being omitted because of the 

 incompleteness of the seasonal representation. 



In 1895 levels were low, unusually so in the spring, and the 

 flood-free intervals of the year were of more than the usual extent. 

 About 10 feet of the total movement in levels (51.9 ft.) is found in 

 the late December rise. If this is excluded, the total movement 

 falls to 42 feet, and the range in levels to 6.5 feet. Under conditions, 



