FEEDING PARAMECIUM KNOWN BACTERIA 175 



might be had if sterile media were inoculated every day with 

 fresh hay, and the cultures obtained used when twenty-four 

 hours old. Experience demonstrated that this method is not 

 reliable. Cultures prepared in this way failed to support the 

 life of Paramecium after a period of four months. Since bacteria 

 are known to vary considerably in their resistance to drying, it 

 is very probable that in the method used, I was actually dealing 

 with a progressively changing bacterial content rather than a 

 fairly constant one, as I at the time supposed. It is, therefore, 

 probable that some such method as described by Jennings is to 

 be preferred for keeping Paramecium in vigorous condition upon 

 a chance mixture of bacteria. It is possible that the determina- 

 tion and selection of the predominant bacteria in an infusion 

 by pouring agar plates every day or two, and seeding infusions 

 with predominating colonies thus obtained, without resorting to 

 first growing them in pure culture, might give satisfactory mix- 

 tures. This method is, however, open to objection. It is at the 

 best tedious and more time-consuming than the familiar way. 

 Moreover, one is not sure that all types of bacteria found in 

 infusions grow equally well on the ordinary media. Considerable 

 study would be necessary to determine this, and additional time 

 consumed in finding a suitable medium for the growth of all types 

 of bacteria. 



The rate of fission is our main index of the metabolic condition 

 of Paramecium. Calkins ('02 a) introduced the method now in 

 use of taking the average di\dsion rate for a given number of 

 days as the basis for conclusions regarding the behavior of this 

 infusorian. The number of days chosen in computing the averages 

 is arbitrary and varied with different investigators. Calkins- 

 preferred the ten-day interval. Woodruff used the five-day in- 

 terval in his investigations. Since, in deahng with the effect of 

 known food, it has seemed desirable to have a record of all fluctua- 

 tions possible and avoid undue error, a three-day interval was 

 chosen for averaging division rates in this study. 



Although it is true that the division rate is the only visible 

 index of the progress of metabolism in Paramecium, and it must 

 serve as the basis for all data bearing on the subject, the phe- 



