PRIMARY PRODUCTIOX 527 



terns. It is this essential nonlinearity which makes it necessary to 

 set up a theoretical picture before the data can be analyzed quan- 

 titatively, and which limits the value of data collected rather at 

 random, for example, during the grid type of survey when changes 

 in the population are occurring. 



Because of these difficulties with time changes, the so-called 

 steady state conditions are often considered theoretically, and 

 looked for in natural events when the populations might be ex- 

 pected to be stationary or to change only slowly with time. Yet 

 situations where this could be thought to hold are often masked 

 by the great variability found in the plankton samples in an area 

 which hydrographically appears homogeneous. 



Since this variability is a major difficulty, I shall try to show 

 some of the problems of sampling and of interpretation which it 

 raises. By thinking of phytoplankton in terms of its carbon con- 

 tent, we imply that it is essentially passive and uniformly com- 

 posed, and so the question arises as to whether its variability can 

 be explained by the variations in its environment. These can be 

 studied by starting from the hypothetical steady state conditions 

 of the deterministic model and producing simplified forms of 

 stochastic processes by introducing natural or random fluctuations. 

 Thus one of the main controlling factors is the variation in daily 

 radiation. I shall assume that the mean photosynthesis per unit 

 of population in the euphotic zone is proportional to the total 

 radiation. This is rather extreme since there will be damping effects 

 due to near-surface inhibition by strong light and to changes in 

 the depth of the euphotic zone due to the size of the population. 

 However, it is a first approximation, and Fig. 3 shows the effects 

 using light data from Aberdeen in May and June, 1957, expressed 

 here as deviations from the average over the period considered. 

 It can be seen that quite large day-to-day fluctuations during the 

 first half do not produce very large changes from the mean in the 

 plant population. It is only in the second half, when there are 

 periods of consistently high and then low intensities, that the popu- 

 lation changes significantly. Thus a week with good sunshine 

 increases the population by a factor of three. 



This also shows the complex relations between light and pro- 



