456 



W. T. Edmondson 



tended to coincide with periods of high illumination ; nevertheless, the coincidence 

 is not perfect, and many variations in photosynthesis are not related to light (Fig. 1). 

 Obviously, the size of the population must be taken into account, since it showed 

 large variations. The rate of photosynthesis per unit volume of water is the product 

 of the population and the rate per unit of population. Even more pertinent to an 

 explanation of the changes in photosynthesis is the amount of chlorophyll in the 

 population (Fig. 1). After fertilization the population of phytoplankton increased 

 greatly, and with it, chlorophyll. It is seen that the large amount of chlorophyll 

 present in Tanks 3 and 4 during the period of time centred around July 20 partly 



CHLOROPHYLL //qm/l 



Fig. 3. Photosynthesis per unit of light as a function of chlorophyll concentration. The rates of 



photosynthesis shown in Fig. 2 have been divided by the mean light income (cal/cm^/day) during the 



period of measurement, and the values multiplied by 1000. 



compensated for the much lower light intensity relative to the period around July 13. 

 The generally higher level of photosynthesis in Tanks 3 and 4 is matched by generally 

 higher concentration of chlorophyll. The chlorophyll concentration in Tank 2 

 achieved a value only slightly more than that obtained in the 1946 experiment with 

 the same amount of fertihzer in an unshaded tank, (erroneously recorded as mg/1 

 in the 1947 paper). 



The relationships just discussed are more clearly visuaUzed if reference is made to 

 correlation graphs (Figs. 2, 3, 4), in which consideration is Hmited to the 90 measure- 

 ments made in the period before the re-fertiHzation on August 11. In the discussion 

 which follows, the conventional correlation coefficient is designated as r^g. 



