LONGITUDE OF TORONTO OBSKRVATORY. 29 



Reduction of the Observations. 



Ill reducing the observations, ProC. McLeod gave equal weight to all stars, whilst I 

 einploj'ed weights obtained from the formula used in the redui^tiou of the observations 

 in the determination of the ditfereiue of longitude l)etwi'fu Cambridge and Montreal, in 

 till' paper already referred to. As there were three exchanges of time with the observers 

 at Ihi'ir own stations, and three after exchange of observers, any error due to the ditterent 

 methods of reduction will be eliminated. 



PART II. 



Instruments at Montreal. 



The transit insirument is by Jones & Son, London, and has recently been refitted with 

 bearings, pivots, and micrometer. 



Tlie striding- level is by Fauth & Co., Washington. The scale value used in the 

 reduction of these observations was 0M4r). 



The time-piece is the " Blackmail " mean time clock, by Howard. 



The chronograph used in this work was of the usual American single-pen cylinder 

 form. It was loaned to the Observatory, by the United States Naval Observatory, 

 through the kindness of the Superintendent, Admiral Shufeldt. 



For further details as to instrumental outfit, reference is made to the Transactions of 

 the Royal Society of Canada, Vol. iii. Sec. iii. pp. 112-114. 



Star Places. 



The apparent right ascensions of the stars observed were taken from the " Berliner 

 Astronomisches .Tahrbuch fur 1883," excepting three stars observed at Montreal, two from 

 tht^ Ameri(\an Eidiemeris and one from the Nairtical Almanac. In Table II, the former 

 are marked l)y an asterisk (*), and the latt( r by a dagger (f). The right ascensions of 

 these stars have Ijeen reduced to the Berlin system. 



Computation of Clock Errors. 



The method of reduiing the observations was in effect as follows: — The obs(»rved 

 times of transit were corrected for level and for assumed approximate values of the 

 azimuth, (^oUimation and clock errors. The corrections to these assumed values were 

 then obtained by the method of least squares, giving each observation full weight. In 

 this reduction the collimation error was assumed to be constant throughout the work of 

 a single night, and on no night have the residuals after réduction thrown iiny doulit on 

 the validity of this assumption. Difi'ereut values of the azimuth error were assumed for 



