BIPAKENTAL INHERITANCE IN PARAMECIUM 439 



the matter in such a way as to exclude even this remote possi- 

 bihty. To do this, we may exclude from consideration all the 

 pairs that died before the end of the first period. If the deaths 

 do not occur within seventeen days of the time the common hand- 

 ling occurred, there is evidently no danger that they were due to 

 this handling; particularly in animals that produce a new genera- 

 tion every twenty-four hours. There were 8 complete pairs that 

 died during these first seventeen days. Excluding these com- 

 pletely, we have 466 Hues (233 pairs) instead of 482 Unes (241 

 pairs) to consider. If now we examine the mortality in the later 

 periods among these 466 lines, omitting the 8 pairs that died during 

 the first period, we have the results given in the second half of 

 table 44. 



As the table shows, essentially the same relations are shown 

 when we proceed in this manner, as when we include all cases. 

 In every period the number of pairs whose members have the same 

 fate is greater than would be probable if the distribution of 

 deaths had no relation to the pairing. The excess becomes greater 

 and greater in the later periods, showing that there is a tendency 

 for the mates of those dead to die, and of those alive to live. The 

 only difference made by omitting the eight pairs of the first period 

 is to make the numerical expression of the improbability less. 

 Yet even thus it rises to 247.6 against 1, in the last period. 



Thus there is absolutely no escape from the conclusion that 

 there is something in the pairing which tends to make the two 

 members of the pairs alike in their fate. 



Comparative rate of reproduction in the members of pairs 



We may next examine the rate of reproduction in the members 

 of pairs, as shown in table 51, in order to determine whether the 

 members of pairs are more alike or less alike in this matter than 

 would be expected. 



The best way to do this is to determine the coefficients of 

 correlation between the members of the pairs, for the di\'erse 

 periods of the ex])erinient. It may possibly be of interest, how- 

 ever, to first examine directly for some typical period the question 



THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 14, NO. 3 



