176 



KOGEES AND McLEOD ON THE 



weights have been obtained by tombiuiug the probable errors of all the quantities upon 

 which the several values of JA depend. The probable error of the clock correction at 

 Montreal, on June 20th, was large, on account of the small number of stars observed and 

 these being all in one position of the instrument. The cause of the différence between 

 the armature and transmission times in the first and second series has already been 

 referred to. No account has been taken of this in obtainiuff the final value of JA. 



Table VIII. 



The probable errors of the values of -JA from "both series" and from "all sources" 

 given above are those which resvilt from the residuals for each night. The probable error 

 of the " final value of -JA " which results from combining the probable errors of the 

 separate values of JA on each night is + -OOY. 



