OBSERVATIONS WITH THE ZENITH TELESCOPE. 



319 



As neither of these conditions have been met, it appears quite doubtful whether the 

 method pursued has been entirely successful in eliminating the latitude variation. 



At all events it has not been thought advisable to introduce any correction on 

 account of aberration. The foregoing results therefore, as to latitude, depend upon 

 Struve's value of this constant. 



THIRD SERIES. 



October tO, 1892, to December 27, 1898. 



Before beginning this series some changes were made in the building and instrument, 

 the most important being the following : 



The width of the observing slit was increased from 20 to 36 inches. It was hoped 

 that this might reduce the liability to disturbance in the way of irregular refraction. 

 The instrument was fitted with two latitude levels to replace the single one before used. 

 The latter had deteriorated to such an extent as to be worthless for refined determination 

 of latitude. The new levels were furnished by (). Saegmiiller, of Washington, and 

 proved superior to any heretofore employed. 



The Level Values. 



The method of determining these, with an example illustrating the process, will be 

 found in the publication of results for 1894, '95.* That series being a continuation of 

 this one, no change in this particular occurred. 



Six different determinations were made as follows. The upper level is called A, the 

 lower B : 



1.335 



B 



1.006 



Thcrmom. 62 



1.305 





1.058 



24 



1.870 





1.024 



44 



1.175 





0.095 



G7 



1.307 





1.044 



50 



1.273 





1.037 



30 



1802 September 30-October 1 

 December 89-December 81 



1893 March 30-Murch 31 

 Juno 85-June 20 

 October 4-Octobei- 12 



1894 January 7-January 8 



The values employed are as follows: 



1892 October -1893 February 



1893 March - June 

 July - October 

 November- December 



In observing care was taken to keep the level correction small, in order to minimize 

 the effect of any uncertainty as to the true value. 



* Vol. XX, TRANSACTIONS OF Tins Amktctcan Philosophical Soctkty, Article iii. 



A 1.3(15 



/; 1.082 



1.272 



1.010 



1.238 



1.030 



1.280 



1.040 



