ENDOMIXIS IN PARAMECIUM AURELIA. 44! 



largely an arbitrary one and that the ideal graph would present 

 the momentary changes in the metabolism of the cell. Data 

 for such a curve being absolutely impossible to secure, it might 

 seem at first glance that the daily record of division would 

 approach most nearly to this ideal condition. As a matter 

 of fact, the twenty-four-hour period is as arbitrary as the five- 

 day period when it is considered that this is a long period when 

 compared with the metabolic changes in the cell and that the 

 daily record, made at approximately u A.M., would merely 

 give the divisions actually completed during the previous twenty- 

 four hours. For example, let us assume that, at the time of 

 isolation, two animals are present, representing one division 

 during the previous twenty-four hours. The record for that 

 day is one division. One animal is then isolated and it divides 

 within an hour and each of the resulting cells again divide 

 twice before the next isolation. The record for this second 

 day is three divisions, thus the record for the two days shows 

 a different division rate for each day, i. e., one division against 

 three divisions, whereas a more true, but not a perfect, picture 

 of the state of affairs is given by the statement that four divisions 

 occurred in forty-eight hours. One might follow this argument 

 to its logical conclusion and assume that the best method of 

 presentation would be' to average for considerable periods, e. g., 

 10 or 30 days, but this obviously would tend to obliterate any 

 fluctuations in the rate which are not of relatively long duration. 

 The adoption of the five-day period was made in recognition of 

 both of these contingencies, and it was of a duration particularly 

 well suited to show the effect of the process on the reproductive 

 rate, because the process extends over about nine cell divisions 

 or a period of about six days. Consequently the effect of the 

 process makes itself evident in the five-day plot. Certain 

 apparent irregularities in the coincidence of the phenomena 

 are, from an actual study of all the data at hand, clearly due 

 to the fact that the five-day period is not ideal." 



EXPERIMENTS SERIES I. 



The experiments of Series I. may be outlined as follows: 

 A. Study of the periodicity of rhythms and endomixis in 



