800 



THE SAY RE OBSERVATORY. 



The tube had been twice refilled in the interval. Whether from this or from some 

 other cause there had been a change of curvature. 



Except for those observations near the beginning of the period under consideration 

 and those not far from 1888, it is obviously impossible to decide what value of the level 

 ought to be employed. The safest plan appeared to be to exclude from the final results 

 all those eases in which this correction was large. The limit of 1".5 for the older 

 observations and 2".0 for the later ones was adopted. A few values were retained where 

 these limits were slightly exceeded. 



The Micrometer. 



The value of one revolution of the screw employed during the different years, with 

 the method of its determination, is as follows: 



1870-78 50.450 ±.001) Nine elongations of 51 Cephei and J CJrsse Minoris. 



1885-80 50.51',?:! ± .0084 From the Latitude Observations. 



1888 60.B688 One elongation each of 51 Cephei and A Ursse Minoris. 



The observation of the old list in 1890 was carried on in connection with the new 

 one. The details as to Micrometer and Level will be found in connection with the 

 Second Scries. 



Method of Observing. 



Previous to 1889 the Micrometer was provided with only one movable thread. One 

 bisection of the star was made, the effort being to have the thread accurately on the star 

 at the instant when the latter crossed the meridian as shown by the clock. This would 

 not usually coincide with the instant of crossing the middle thread of the reticle, as it 

 was not possible to adjust the instrument with great precision in this respect. The out- 

 standing difference, however, was of little importance. The (dock correction was deter- 

 mined with a small transit instrument and was always known with the requisite 

 precision. 



The level was read once, usually after bisection of the Star. 



A. preliminary reduction of the observations up to dune 20, 1877, seemed to show a 

 tendency of the bubble occasionally to cling to the lube, thus producing fictitious level 

 readings. After this time until the end of the work in 1878, an effort was made to over- 

 come this difficulty, or to obtain evidence tending to show whether or not the readings 

 were genuine, in the following manner : Whenever there was sufficient time between the 

 stars two or more readings of the level were taken, the instrument being disturbed by 



