EHODA.] METHODS FOR DETERMINING ELEVATIONS. 491 



theuce to station 34:, distant G.G5 miles, a fall of 1,161 feet, the mean of 

 five determinations, range 23 feet. 



This completes all the well-determined chains. Other stations on 

 which barometric readings had been taken, were connected with as 

 many points in the main lines as possible, and these being reduced to a 

 common point a mean was taken. Such points are the following: 

 Sultau Mountain to station 10, a fall of 223 feet, the mean of ele^■en 

 determinations, range 76 feet ; IJncompahgrePeak to station 5, a fall of 

 1,498 feet, the mean of ten determinations, range 85 feet. ITncompah- 

 gre Peak to station 11, a fall of 3,624 feet, the mean of eight deter- 

 minations, range 111 feet; Sultan Mountain to station 51, a fall of 835 

 feet, the mean of three determinations, range 75 feet; Sultau Mountain 

 to station 48, a fall of 1,061 feet, the mean of six determinations, range 

 59 feet ; Handie's Peak to station 13, a tail of 1,175 feet, mean of fore 

 and back sights, range 6 feet. 



Besides these there are two which depend on single determinations : 

 First, from Sultan Mountain to the point in Baker's Park where the road 

 crosses Cement Creek in Silverton, distant three miles, a fall of 

 3,961 feet. Second, from Mount Sneffels to station 32, which is obtained 

 from sights to a common point between them, distant from Mount 

 Sneffeis 2.04 miles, and from station 32 3.75 miles, the fall is 5,050 feet. 

 This difference of level is checked by sights to distant points to the 

 south of station 32. These two cases are admitted, because the distances 

 were so short as to preclude the i>ossibilit3^ of any considerable error. 



From all these results a table was made out, showing the heights of 

 each, station above or below a common datum-point. Sultau Mountain 

 was selected as the datum point, from its central location, and also from 

 the fact that it was situated on the great central chain of levels at its 

 junction with the two principal sub-lines. 



A second column was added, giving the height of each station as deter- 

 mined by the single barometric reading taken thereon. A third column 

 was made out from the first two by adding the number in the first 

 column to the one in the second when preceded by the minns sign, and 

 by subtracting it when plus. This column represents the elevations 

 above sea-level of Sultan Mountain, as determined from the barometric 

 readings at the several stations. It will be seen that the twenty-three 

 results have a range of 203 feet. A mean of all these was assumed as 

 the true height of Sultan Mountain, and by reversing the previous 

 jjrocess and adding the plus differences of height in the first column 

 and subtracting the minus, a fourth column was obtained, giving the 

 elevation of each station as reduced, from the mean of the twenty-three 

 readings. A fifth column was added, giving the date of each reading 

 on the different stations. From tbis it will be seen that the observa- 

 tions extend from August 1 to October 6, more than two months. 



By examining the table carefully, it will be seen that nearly all the 

 earlier readings give heights above the mean, and the later below it. 

 Whether tbis is merely accidental or due to some physical law, I cannot 

 tell. It will be seen that several of those stations, whose height relative 

 to the rest has been well determined, do not appear in the table. This 

 is due to the fact that at those stations, either from storms or other 

 causes, we failed to get barometric readings. 



