524 CYCLKS OF ORGANIC AND IXORCAXIC SUBSTANCES 



artificialities of laboratory experiments, the experimental results 

 cannot always be applied straightforwardly to observations in the 

 sea. An example of this is that it is very easy to demonstrate the 

 slowing down of plant growth due to nutrient deficiency in a cul- 

 ture flask, but this effect has not yet been observed directly in the 

 sea (except for som.e recent preliminary results by Dr. Ryther). 

 Again, the experimentally determined filtering rates of zooplank- 

 ton show a wide range of values (Gushing, 1958) which may be 

 due to variations in experimental technique (Gushing, 1959b). Thus 

 it is necessary to check the values by seeing if, in their natural 

 en\'ironment, the filtering rates will provide the animals with 

 sufficient food to meet their respiratory needs (Gonover, 1956). 



In this way difficulties arise in comparing the simple and logical 

 structure of an experiment with the often apparently chaotic 

 sequences of data from the sea. This may lead to a certain diver- 

 gence between the two approaches with the laboratory experiments 

 regarded as the source of hypotheses and the observations gi\-ing 

 merely a broad descriptive indication of the probable consequences. 

 But because of the difiiculties inherent in the methods of sampling 

 and in the laboratory experiments, no hypothesis is really accept- 

 able until there is good evidence for the particular causal relations 

 operating in the sea. 



Once more, then, from this point of view there is the need to 

 de^4se programs which will provide data to test critically particu- 

 lar hypotheses. At the same time such detailed studies can pro- 

 vide ideas about possible features of plant behavior which might 

 not be expected eithei from broad surveys or from work with 

 cultures. The general prol)lem in finding a way to analyze such 

 data is that no particular feature can really be separated from the 

 rest, and one has to consider several possible interrelations at the 

 same time. I should like to illustrate now some of the com_plIca- 

 tions that arise in a detailed study of observations by using a 

 theoretical model of the production cycle which I developed for 

 the North Sea (Steele, 1958). To construct it I used estimates of 

 photosynthesis, respiration, grazing rates, and so on, all of which, 

 as I have said, are rather dubious, so that the model demonstrates 

 what the questions are rather than providing answers. 



