FEEDING PARAMECIUM KNOWN BACTERIA 155 



with either of the artificial mixtures, the daily rate for M averaging 

 1.44-4 for ninety-two days. But the differences between these 

 mixtures are not sufficiently great to have any meaning, since 

 we find the significance factors as follows: M and A'C, 2.880; 

 M and A'B'C, 0.704. If we consider the control mixture M only 

 for the first three periods, during which time it supported a 

 higher metabolic rate in Paramecium, we find no ground for 

 changing the conclusions already given. For these three periods 

 M was superior to C, while M, A'C, and A'B'C did not show 

 sufF.cient difference to give a significance factor equal to 3 in 

 any case. 



These figures thus make it clear that the mixtures of bacteria 

 maintain a higher rate of metabolism in Paramecium than do 

 the pure cultures of the bacteria comprising the mixtures. There 

 appear to be differences in the mixtures, for the daily rate of 

 division would indicate that A'C was best, then comes A'B'C, 

 and M last. But the biometrical tests applied showed the dif- 

 ferences to be too small to warrant any belief in the marked 

 superiority of one mixture over the others. 



The time during which the various cultures of the bacteria 

 A', B', and C were used for feeding was divided into five periods, 

 because of the necessity for occasional washing of the animals 

 and because, for the purposes of this investigation, it seemed 

 wise to avoid experimental work during endomixis. The phenom- 

 ena of endomixis are accompanied by a marked slowing of the 

 division rate, and it was thought that the effect of the different 

 kinds of food would not be so evident at these times. The oc- 

 currence of fragmentation of the macronucleus as shown in 

 stained specimens was taken as evidence of endomixis, and this 

 was determined as taking place in this strain of Paramecium 

 every thirty to thirty-five days. 



It was found to be well-nigh impossible to revive the animals 

 after the first endomixis period, which began September 28, 1920. 

 Vigorous, rapidly dividing animals were not secured until October 

 10th. I have been at a loss to account for this condition of low 

 vitality, save that it was in some way the result of environment. 

 The animals passed through this endomixis in the control mixture 



