528 



CYCLES OF or(;axic and inorganic substances 



/] 



w 



nfl 



Ln 



L^^ 



ooe 



■0 02 S 



o 02 ■ 



Phoiphorus 



2ooplankton 



Days 



Fig 3. The effects of natural light data introduced into the theoretical model. 



tluction, for although the production per unit of population on any 

 (lay depends mainly on the light during that day, the total popu- 

 lation itself depends on the light during, say, the previous week. 

 Thus a C^-* in siiii measurement is a complicated function of light 

 over a fairly long period of time. Again the probability of the 

 population varying with light depends on the nonrandomness of 

 runs of bright or dark days. Even with British weather such runs 

 of fi\'e or more days do occur quite frequently. This in turn tends 

 to upset the use of average \'alues such as Kimball's tables. Figure 



