Studies on the Ecology of a Marine Spirillum 563 



Do 



Sk = Ks (2) 



,U„. - Do 



Assessing Dc by lowering Sr, theoretically represents a new 

 means of evaluating Ks. As the experiments show, however, the 

 exact position of D. is difficult to determine, due to large fluctua- 

 tions in the cell count when wash-out approaches. 



From formula 2) the theoretical value of Sk, causing wash- 

 out at the given dilution rate, is 0.06 mg asparagine-N/L. The 

 experimental value, however, must be somewhat above 0.14 mg 

 asparagine-N/L ( Fig. 1 ) . In other words, as regards a decreasing 

 Su, wash-out appears earlier than one would expect from the 

 growth constants measured in batch culture. According to Fer- 

 ret (12), there may be several reasons for this deviation. Powell 

 (13) supposes a decrease of viability with decreasing substrate 

 concentration and dilution rate, a phenomenon which contributes 

 to Ferret's growth rate hysteresis. In general, interpretation of 

 results obtained with the chemostat at low values of Sr as well as 

 of D involves inherent difficulties as indicated by the frequency 

 function of generation time. Moreover, at growth rates as low 

 as /^m/2, the cell number may be influenced by an inconstancy 

 of the yield as observed by Holme (6) and Herbert (5). 



When the substrate concentration was shifted from 0.28 to 

 0.14 mg asparagine-N/L, the cell count did not decrease in pro-> 

 portion to the dilution rate, thus still indicating growth. Never- 

 theless after 60 hours (ca. 11 volume changes of the culture 

 vessel), cell count was practically zero. Large fluctuations of the 

 cell count below 5 000/ml must also be attributed to the increas- 

 ing statistical error. 



Contois (3) in a study of the relation between specific 

 growth rate and population density showed that D. was inde- 

 pendent of Sh, thus making Ks a function of Sr. Although D. was 

 not determined with numerical exactitude in the present work, the 

 dependence of its decrease on Sr was clearly obser\ ed. 



The threshold concentrations of the growth-limiting nutrient, 

 as here reported, naturally depend on the given dilution rate. If 

 the dilution rate, beginning at D., is lowered at a constant Sr, 

 the substrate concentration in the effluent (S) decreases. This 



