STARS IN COMA BERENICES. 403 



peat the operation as for the East Image. Take the mean of 

 the readings on the scale, and subtract from it the mean of the 

 readings oh both images. The difference, divided by 2, is the 

 distance in milhmeters of the mean position of the star beyond 

 the given hne on the scale. It is designated in the following 

 by ^ in. Measure in this way all the stars, beginning with the 

 central. Read the temperature. Remeasure all the stars in 

 the inverse order, with the micrometer head set now at half a 

 revolution from its previous position, in order to eliminate peri- 

 odic errors of the screw. Read the runs, circles, and temper- 

 ature. 



In this way, on one day, all the stars were measured in one 

 position of the plate. Two observers were always engaged on 

 the work, each one reading all. the stars, the runs, and the 

 circles. After completing the measures in one position, the 

 plate is rotated through 90°, and the process is repeated. It is 

 then evident, that if the first position gave differences in right as- 

 cension, the second would give differences in declination. Since 

 the two images are separated on the plate by about a millimeter 

 in right ascension, it was in general necessary to use two lines on 

 the scale when measuring that coordinate, while for the other only 

 one was required. In all other respects the measures in both 

 positions are entirely similar. To reduce personality, observa- 

 tions were made with the plate respectively 180° and 270° from 

 its original position, care being taken that the same pair of ob- 

 servers should always read one coordinate in both directions. 

 The greater part of the systematic error, due to the difficulty 

 of judging the center in hazy images, cannot be eliminated by 

 this method, however. The only way to obviate its effect, is to 

 multiply the number of the plates, if that is possible. 



In Tables II and III are recorded all the observational data 

 from which the succeeding reductions are made. Table II 

 gives the daily record : It shows the date of measuring the 

 plate, the runs in millimeters on 10 mm. spaces, the circle 

 readings, the mean temperature of the morning, the position of 

 the plate, and the initials of the observers, Schlesinger, Hays 

 or Kretz. The runs as here given are the mean of the two 



(63) 



