Mr. W. Ellis. On the Simultaneity of 



stations the times are distinctly greater and at others distinctly less 

 than the average of times for all stations. Either the magnetic im- 

 pulse is really retarded at some stations and accelerated at others, 

 as referred to the mean, or there exists at individual stations some 

 systematic error, mechanical or otherwise, which affects the times in 

 one particular way. That such error should exist is not unlikely 

 since it is difficult to make accurate, beyond a certain point, a time 

 scale so contracted, and difficult also to measure correctly the photo- 

 graphic trace, besides which the desirability of securing such extreme 

 accuracy has not been before made apparent. 



It happens that on five of the selected days, Nos. 6, 7, 8, 9, and 10 r 

 corresponding records were found in the magnetic section of the 

 'Mission Scientifique du Cap Horn,' 18821883. Here No. 6 is 

 said to be " une diminution brusque," No. 7 " un mouvement tres 

 rapide," and No. 9 is said to commence " brusquement." Engraved 

 copies of the registers on a small scale are given for Nos. 8, 9, and 10, 

 showing the movements to have been of the character described. 

 This affords five comparisons. The times are given only to the 

 nearest five minutes, and the mean of the five resulting deviations^ 

 which range from -f 7'lm. to 8'6m., is + I'l m. There were in 

 addition six other instances of sudden movement contained in the 

 Cape Horn series, not included in our selected list, for which corre- 

 sponding movements were readily found in the Greenwich records. 

 These movements are spoken of as having at Cape Horn the same 

 abrupt character as those already mentioned. The separate values 

 of deviation referred (by comparison through Greenwich) to the 

 general mean of Table I range from -f-9'2m. to 10'8 m., the mean 

 being 2'1 m. Or, combining the two sets, we have from eleven com- 

 parisons a mean deviation of 0'7 m. 



The mean deviation of time at each station from the general mean 

 for all stations (the means of the columns of differences in Table I) 

 and the difference between the greatest and least values at each 

 station, adding the result for Cap.e Horn for eleven days just found,, 

 are as follows : 



