112 C, M. CHILD 



increase in normal head-frequency in c. In all cases the results 

 of long periods in KCN are in marked contrast to those of short 

 periods both as regards the greater effectiveness of long periods 

 in shifting head-frequency downward in a, and their lesser effect- 

 iveness in increasing normal head-frequency in c. The reason 

 for this difference will appear below. 



Certain differences of minor importance between the differ- 

 ent long period series appear. First, the decrease in normal 

 head-frequency produced by KCN m/200000 in a is consider- 

 ably greater in table 5, III a, than in table 3, II a and III a, 

 or in table 4, III a. Again, the upward shift in head-frequency 

 in c is much greater in table 5, III c, and in table 3, II c and III 

 c than in table 4, III c. These differences are due to factors 

 not fully controlled in the experiments. My experience with 

 hundreds of such series has shown that inaccuracies in the 

 level of section probably play a larger part in determining such 

 differences than any other factor. Each series is prepared at 

 a single sitting, but I have found that a slightly different habit 

 may be established in different series. In one series, for example, 

 the c pieces may average a little larger than the a-pieces or vice 

 versa, or the levels of section may differ slightly in different 

 series, so that the c-pieces in one series include a part of the 

 second zooid more frequently than in another. On the other 

 hand, the sensitiveness of the material to en\ironmental con- 

 ditions, which will become increasingly apparent in following 

 papers leaves no doubt that slight differences in temperature, 

 constitution of water or nutritive condition of the worms also 

 play a part in producing such differences in different series. 

 But these differences are not sufficient to obscure the character- 

 istic features and it is a striking fact that different series, of 

 experiments, performed at different seasons of the year with 

 worms which have been kept in the laboratory and have re- 

 ceived a food (beef liver) different from the natural food for 

 different lengths of time show so high a degree of similarity. 



In series 559, table 6, the effects of periods of different length 

 in KCN m/100000 are compared. The worms used in this 

 series were considerably longer than those of the other series 



