79 



though it is frequently done. If nearly the same number of stars are observed every 

 night under the same conditions, I should prefer to determine the probable error of 

 the final result after the manner of Mr. John H. Clark. 



Let- 

 Probable error of one pair of stars, including constant errors 



of zenith-telescope observations = f 



Probable error of one observation rr e 



Number of pairs used at the station m 



Number of observations . 



then probable error of the final result, 



The formula shows that if the stars used are not very good, it is then better to 

 select a larger number of pairs of stars, giving the probability that the final result will 

 be more independent of the declinations. 



From all the observations of pairs of stars observed on three or more nights, I 

 find the probable error of one observation 



and that of the final result, 



Taking the value for e f , found by Maj. C. B. Comstock, of the United States Lake 

 Survey, for stars taken from Professor Safford's Catalogue for 98! Stars, o".53, 

 the probable uncertainty of the final result will be, rbo".o82. 



Resulting Astronomical Co-ordinates for the Astronomical Monument at Colorado Springs, 

 Colorado Territory, using, for the longitude of Washington and Salt Lake, the same data 

 as in Clark's report. 



In time. In arc. 



h. m. s. 



Longitude . . . i 51 4.888 27 46 1 3.30 west of Washington. 



6 59 17.008 104 49 15.10 west of Greenwich. 



Latitude (north) ..... 38 49 41.67 o".o35 



Respectfully submitted. 



DR. F. KAMPF, 

 Civilian Astronomical Assistant. 

 First Lieut. GrEO. M. WHEELER, 



Corps of Engineers, in charge. 



