ENDOMIXIS IN PARAMAECIUM AURELIA 39 



part, when daily changes were made, the rate was 2, showing a marked 

 increase. During the former period endomixis occurred on an average 

 of once every twenty-five generations, while during the second part 

 its occurrence was reduced to once every ninety-one (two) generations, 4 

 showing the marked influence of purity of culture medium in reducing 

 the frequency of endomixis. 5 



Experiment 2 (figs. II and Ha) 



In this experiment line II was branched off from I on 2/8 and car- 

 ried through one hundred and six generations with only occasional 

 changes to fresh slides, two or three drops of fresh culture however, 

 being added daily. From 4/21 to 5/6 a 0.5 per cent solution of urea 

 in distilled water was used as a culture medium in an effort to induce 

 endomixis. This failing, the line was returned to beef extract on the 

 latter date. On 5/9 line Ha was branched off from II and carried 

 until 6/19 with fresh slide changes daily. The details of the experi- 

 ment are shown in the accompanying graphs, which show that in line 

 II endomixis occurred on the average of once in every thirty-five gen- 

 erations 6 while in line Ila it occurred once in seventy-eight generations, 

 an increase of forty-three generations. 



It is possible that the change to 0.5 per cent urea solution followed 

 by a re-transfer to beef extract may have induced the long interval of 

 65 generations between the endomixis of 3/5 and 5/13 in line II. This 

 suggestion is borne out by the fact that in line IV, which was branched 

 off from II on 4/13 and carried without change of slide until 5/12, 

 endomixis occurred on 5/1 forty-five generations after the last 

 precedent one in line II, which appeared on 3/5. It is also supported 

 by the results of Calkins ('02-04) who showed that various stimuli 

 (change of food, temperature, shaking and chemicals) tend to increase 

 the division rate. 



Experiment 3 (fig. Ill) 



On 5/9 line III was branched off from II and carried with daily 

 slide changes until 6/19, endomixis occurring on 5/13 and 6/13, with an 

 intervening period of sixty-four generations, an increase of twenty-nine 

 over thirty-five, the average period in II, the parent line. 



4 Counting from the last endomixis in line I before la was branched off from 

 the latter. 



5 My notes show an endomixis in line I on 3/16. This is in all probability an 

 error because of its close proximity to those on 3/8 and 3/23, and because the 

 notes give no record of multinuclearity on either 3/15 or 3/17. I have accordingly 

 rejected it in computing the average period of 21 generations between endomixes 

 in this line. If it were counted, however, it would reduce the average to less 

 than 21 and make the contrast even greater between the average period in line 

 I and the period in la. 



G Counting from the appearance of the last endomixis in line I from which II 

 was branched off. 



