188 CARNEGIE INSTITUTION OF WASHINGTON. 



expected to result from a larger number of observations taken under 

 the same conditions at one observatory. A supplementary list of 

 2,600 stars of T'fO to TfO, lying within the same limits of declination, 

 has been supplied to Professor Comstock. The stars of this list will 

 be observed at Madison as far as practicable. 



OBSERVATIONS. 



During the past year 12,405 observations have been taken with 

 the meridian-circle telescope. These have been distributed among 

 the observers as follows: B. Boss 3,283, H. Raymond 339, A. J. Roy 

 4,113, and W. B. Varnum 4,670. As usual, four microscopes have 

 been recorded at each setting in zenith distance. Indicating the 

 circle nearest the clamp by A, that opposite the clamp by B, the 

 readings have been distributed CI E and CI W as follows: AE 3,184, 

 BE 3,032, AW 3,961, BW 2,228. The distribution among the circle- 

 readers follows: S. B. Grant 5,179, A. R. Guy 782, H. Henderson 

 292, H. Jenkins 1,645, and H. Raymond 4,499. The plan of obser- 

 vation has been the same as in previous years. 



The flooding of the Hudson River Valley in March, because of its 

 interference with the lighting system of Albany, caused the loss of 

 some observing nights, and in a part of April and May still more 

 time was lost during an effort to seal the Riefler clock. 



REDUCTION OF OBSERVATIONS. 



The reduction of current transits to the mean wire have been kept 

 nearly up to date and checked, and the wire intervals for each 

 observer drawn off. The material for magnitude equation is pre- 

 pared currently in ready form for final evaluation. The apparent 

 place reductions have also been duphcated nearly to date. 



In working up the contacts of the micrometer wire with the middle 

 wire Mr. Roy noted a distinct variation dependent upon tempera- 

 ture. By adopting a term 34^0104— 0! 00033 (T-0°) the observed 

 values were fitted very closely, with a probable error of ±0! 0017 for 

 a single setting. The difference between day and night readings, 

 — 0y0014, though about five times the probable error of the determi- 

 nation, has been neglected. 



The effect of the screen used in reducing the magnitudes of bright 

 stars to approximately a uniform magnitude was found to be 3V2. 

 Although the same screen was employed in the San Luis observations 

 its effect at that station was 3Y8. The effect would appear to be real. 



Since the last report the definitive azimuths for the San Luis 

 observations have been computed and adopted. The accompanying 

 table gives the resulting mean readings of the mire-mark for each 

 month arranged in 6-hour groups according to mean time. The 

 colunm at the extreme right represents the straight mean of the 



