480 



LORANDE LOSS WOODRUFF AND RH. ERDMANN 



view of its nuclear state. By this method the process was dis- 

 covered at the 4065th, 4189th, 4237th, 4315th generation (cf. 

 table 2). Text figure 17 gives the graph of the division rate of 

 this line, plotted the same as that already presented for Line 

 VI, and shows the periods in which the generations involving 

 nuclear reorganization were discovered. Here again in four out 

 of five cases, the process coincides with the low point in division 

 rate, i.e., between two rhythms. The occurrence of the process 

 of the 4315th generation, however, coincides with the early 

 ascending phase of the rhythm and is an exception for which the 

 data afford no evident explanation. 



Text fig. 17 Graph of the rate of division of Line III, subculture IE, averaged 

 for five-day periods. The periods during which the reorganization process oc- 

 curred are indicated by an X. 



The other main lines were carried chiefly for the purpose of 

 affording a sufficient supply of animals in the process and, after 

 the first couple of months, were preserved only when the process 

 was suspected from the appearance of the cells, the rate of divi- 

 sion and the length of time and the number of generations since 

 its last occurrence. Consequently it is unnecessary to consider 

 these lines in detail from the standpoint of the rhythms but 

 simply to emphasize that the evidence derived from them is 

 entirely corroborative of that presented from Lines VI and III 

 (cf. table 3). In our description of the cytological changes in 

 the process it has l^een shown that the phenomenon extends 

 over about nine cell generations, and that at the climax divi- 

 sion is deferred for a period of nearly 36 hours. Therefore, it is 



