These values of e, are therefore slightly worse than those of the preceding set, 

 and still some three times as bad as those of Potsdam I. They seem, however, to be 

 somewhat more free from systematic errors. 



For the differences of each pendulum from the mean, Tables XLI and XLII are, 

 as usual, formed below : 



TABLE XLI. 



TABLE XLII. 



mean 7-7 XlO- 7 sec. 



From these we obtain : 



for the difference M y , 5 = 8- 1 X 10~' sec. 

 M p 7 -6-3X10-' 

 21 -- M /( == 8-6X10-' 



The mean value 7-7xlO~' sec. is more than twice as great as that calculated from 

 the observations of series " C " a result which is probably to be referred to the shaking 

 of the whole hut by the force of the blizzard wind. 



56 



