502 J. A, DAWSON 



For the purpose now of making a comparison of cultures of 

 cannibal and non-cannibal animals, eight new subcultures were 

 begun from these subcultures as follows. On January 22nd 

 from stock slides of four days standing four cannibals were 

 selected — two from subculture 4A to begin subculture 4A-C and 

 4A-1C; one from subculture 4B to begin subculture 4B-C, and 

 one from subculture 3C to begin subculture 3C-C. In all cases 

 as soon as the animal used to begin the respective subcultures 

 had divided twice, four lines in each subculture were estab- 

 lished. Non-cannibal animals, from the same slides from which 

 the cannibals were taken, w^ere selected to form the four lines 

 each of subcultures 4A-N, 4A-1N, 4B-N, and 3C-N. The 

 curve shows that the average division rate for the cannibal 

 cultures (heavy dotted line beginning at B) was higher, 1.40 as 

 compared with .93, than the average division rate for the non- 

 cannibal cultures (continuous line beginning at B) for the first 

 five-day period, but that otherwise little difference can be noted. 



On February 9th (continuous and dotted lines beginning at 

 ])t. C. in fig. 3) two subcultures, 4A-C1 and 4A-N1, were begun 

 from a cannibal and a non-cannibal, respectively, selected on 

 the previous day from a six-day-old stock culture of 4A-C. 

 The difference in division rate here is not so striking as in the 

 earlier part of the experiment, but the cannibal line is slightly 

 higher during the first full five-day period, and this difference 

 increases in the favor of the cannibal line during the rest of the 

 life of the culture. On February 26th these series of subcultures 

 were discontinued. 



The general result of these experiments shows that the effect 

 of cannibalism is to produce an initial higher division rate in the 

 cannibal lines. This higher fission rate may be due merely to 

 the extra nutrition supplied by the ingested animals, in view of 

 the fact that Joukowsky ('98) also found a distinctly higher divi- 

 sion rate in cultures of Pleurotricha fed on the ciliate, Uronema, 

 in hay infusion, than in those kept in infusions of hay, flour, or 

 •albumin alone. Since, in this case, however, animals of the 

 same species formed the food in question, there is a possibility, 

 which should not be overlooked, that a more fundamental meta- 



