112 KOGEES AND McLEOD ON THE 



the A-ahie 4'' 54'" 18~.87, which differs fi'om the détermination through Quebec by 1'13. 

 This discrepancy, notwithstanding the acknowledged imperfection of the old determina- 

 tion, was unexpectedly large and pointed to the necessity of a new determination by 

 direct connection with Cambridge. Professor Pickering very kindly granted the use of 

 such of the apparatus of the Harvard College Observatory as might be necessary in the 

 prosecution of the work, and at his request Professor W. A. Rogers undertook the deter- 

 mination with me. One half the expense of the work was borue by the Harvard College 

 Observatory. 



The Station at Montreal was that occupied by the transit instrument of the McGill 

 College Observatory. It is situated iu rear of the main college building and twenty feet 

 to the west of the meridian passing through the centre of the Blackman telescope pier. 



Instruments at Montreal. 



The transit instrument, made by Jones & Son, London, has been refitted with bear- 

 ings, pivots and micrometer by Mr. Gr. F. Ballou, of Boston, under the superintendence of 

 Professor Rogers. The focal length is 43 in., and clear aperture 32Ô in. The power of the 

 diagonal eyepiece used is 84. The object-glass is believed to be by Dollond. The pivots 

 are of steel shrunk on the original gun-metal. The pivot bearings are of phosphor-bronze 

 and are in the form of segments of a cylinder of the same diameter as the pivots. The 

 diagram exhibits a section of the pivots and bearings to actual size. 



A series of level tests made in the months of April, May and June, 1884, show that 

 the pivots are truly cylindrical in form, but the evidence thus obtained is not quite so 

 conclusive as to their equality. On the opposite page are given the results of these 

 observations, arranged according to the temperature of the air at the time of the test. 

 Each set is obtained from the mean of six level readings. 



The arithmetical mean of the whole 81 sets gives — 0\0017, and the mean tempera- 

 ture is 60'. These results seem to show that this correction is a function of the 

 temperature, and from the construction of the piA^ots it is not unreasonable to sup- 

 pose that such is the case. One pivot is hollow and the other solid. Assuming that 

 the rate of expansion of the solid pivot is that due to gun metal, and that the rate of the 

 hollow end is that due to steel, the change in the correction for inequality of pivots, per 

 degree Fah., would be O'OOIS. Neglecting the observations on April 21st, which do not 

 accord well with the remainder of the series, this correction, as obtained by observation isj 



