328 PROCEEDINGS OP THE AMERICAN ACADEMY 



their observations furnish only values of n and equations of condition 

 (not here given) among the various stars. 



The instrument was not reversed during the present series, but the 

 eye-piece and object-glass were interchanged, September 4, 1883. 

 Before that date the observations were not so accurate as afterwards, 

 owinji to various small causes of error. The instrument was liable to 

 slight strains, until the counterpoises were properly regulated ; this 

 affects the observations up to April 17, 1883. There was a month's 

 interruption after this date, due to illness. The collimation was a 

 little uncertain during the winter; the assistant who determined it by 

 level and nadir met with some difficulty in levelling, and his results 

 were not very accordant. The clock was frequently inaudible ; much 

 of the time an assistant gave me the ticks during transits, and relief was 

 finally obtained by a galvanic connection with a telegraphic sounder. 

 This took place in or about July, 1883. The exchange of object-glass 

 and eyepiece put an end to a trouble with the field illumination, which 

 was due seemingly to a displacement of the prisms in transportation. 



Very few observations are made upon less than five wires; half- 

 weight * is given to one or two wires ; and in the case of the stars 

 nearest the pole the movable vertical wire is freely used. As its coin- 

 cidence with the fixed wire varies a little with the temperature, it is 

 necessary to determine its coincidence with the central fixed wire very 

 carefully. The reductions to mean place have been effected with the 

 data of the American Ephemeris, — which agree closely with those 

 of the Jahrbuch, — and have been duplicated in many cases, employ- 

 ing Bessel's two sets of formula;. As declinations have not been 

 observed, they have .been pretty carefully computed from modern 

 catalogues, so that their values as employed are usually correct 

 to 1". 



* More frequently the half-vvcijilit is clue to a combination of unfavorable 

 circumstances noted at the time of observation, or in one case to large and 

 irregular changes in the instrumental correction, so that its true value was un- 

 certain. One observation has been rejected for this reason ; but may per- 

 haps prove reducible later. 



