332 



of 1896, our year of fullest observation, by over fivefold. These 

 two years are much alike in that both (cf. PI. X. andXII.) have 

 continuously falling levels from a "June rise" just culminated, 

 with similar rates of decline and much the same flood duration. 

 They differ in the fact that the level of 1898 is on the average 

 4.2 feet above that of 1896. Higher levels with increased res- 

 ervoir action of backwaters thus seem to conduce to greater 

 production in this instance. 



The contrast in the amplitude of the April-May and the 

 June pulses (PI. XII.) is very striking, the former being five- 

 fold greater than the latter. Both occur on declining floods at 

 almost identical levels ( 11.1 and 11.9 ft.). The principal differ- 

 ences in environmental conditions lie, first in the higher tem- 

 peratures (by 20°) in June, and, again, in the duration of the 

 flood on whose decline the pulse appears. The April-May pulse 

 appears at a level which had been exceeded by ten weeks of 

 overflow, while that of the June pulse had been exceeded by 

 only four. Greater time for breeding is thus afforded in the first 

 instance. I am inclined to think that the main factor in this 

 decreased production is to be found in the fundamental mid- 

 summer decrease apparent in most years and localities in our 

 quantitative catches by the silk net. . To this decrease summer 

 heat may be one of the contributing factors. 



From this point onward in our records the variations in 

 amplitude of the fluctuations in plankton production are but 

 slight, and it may seem from the volumetric point of view of 

 little importance. Since, however, they may continue to illus- 

 trate what I have called the cyclic movement in plankton pro- 

 duction I shall endeavor to trace the recurrent pulses wherever 

 they can be found in the data. 



The July pulse has a duration of 21 days, — from the 5th to 

 the 26th, — with a maximum amplitude of .88 cm. 3 per m. 3 on the 

 19th. Its mean falls on the 18th, 34 days after that of the pre- 

 ceding pulse. This is a month of most pronounced fall in river 

 levels. The distance between the extremes (9.4 and 2.5 ft.) 

 and the rate of decline (.25 ft. per day) are unequaled in any 



