TABLE XLV. 



From this, we get the probable errors of a single pendulum comparison : 



M^ 5 = 3-8X10- 7 sec. 



Mp 7 =8-6 



21 -- Mp = 8-8, the mean value of which is 7-1 XlO~ 7 sec. 



The agreement between this figure and the mean value of e b , e-,, and e zl may be 

 taken as an indication that the chronometer " S.C." was behaving satisfactorily during 

 March 2nd and 3rd. 



A comparison of the mean values of this series with the 1910 series shows, however, 

 a lamentable lack of agreement. Thus, the values in the two years for the three 

 pendulums were, respectively : 



Except in the case of pendulum No. 21, the discrepancies are quite outside the 

 limits of probable error, even allowing for the very unsatisfactory observations of 1910. 



60 



