LETTER OF THE GEOLOGIST. XXVII 



hence the local time. By this method the longitude of a point can be 

 determined within 30, 40, or 50 feet at the worst. 



Now, the latitude and longitude of a point being known, those of all 

 other points also are determined, as their distances and directions from 

 the first point are given by the triangulatiou, and the whole system is 

 correctly placed on the earth's surface. 



In the inception of the work in Colorado a base line was measured 

 near Denver, mainly on the track of the Kansas Pacific Eailroad. Two 

 measurements were made of it, with a steel taj^e 100 feet long, under a 

 tension of 16 pounds. The end of each 100 feet was marked by a knife- 

 edge on the railroad track, or on a low stool. The profile of the line was 

 leveled, and the temperature of the tape was constantly measured. 



The results of two measurements, corrected for temperature and slope, 

 are respectively 31,861.304 and 31,863.102 feet, showing a discrepancy 

 of about 1.8 feet, or about ^5^00 o^ the length. It was, of course, cor- 

 rected for error of tape, and reduced to sea-level. 



From this base, triangulatiou was extended and carried into the 

 mountains. Then a second base was measured in San Luis Valley. 

 The methods were the same, and the results better than in the case of 

 the Denver base. Six days were occupied in its measurement and re- 

 measurement. The corrected results are as follows: 28,522.74 and 

 28,522.558 feet, a difference of only -^^^ of a foot in 5^ miles. 



Connecting the triangulatiou from those two bases with one another, 

 the error was found to be only 9 J inches to the mile. This error is the 

 sum of the errors of the measurements of the bases and of the triangu- 

 latiou as brought through nearly 200 miles. The angles were measured 

 with theodolites whose circles were 8 inches in diameter, receding to 

 10'' of arc. All the stations were marked by monuments 5 to 10 feet 

 in height. 



In the whole scheme of triangulatiou of Colorado, 143 complete tri- 

 angles have been measured, with a mean error of closure of 13".3. The 

 errors of measurement of the sides of the triangles will not exceed ^ 

 of a foot per mile, and the error of location of the primary stations is 

 not greater than 25 feet. The area covered in Colorado is about 70,000 

 square miles. 



For the location of the work in latitude and longitude, it was deemed 

 best, on account of the expense of instruments, &c., to ask the Coast 

 Survey, which is completely fitted in this respect for the work, to 

 determine the latitude and longitude of several points for us. This 

 they very kindly consented to do, and established the position of points 

 in Denver, Colorado Springs, and Trinidad, Colo., for us with the greatest 

 accuracy. 



The next point is, how these locations are to be used for the inception 

 of land surveys in isolated mountain valleys, and for the elimination of 

 errors in the work of these surveys. 



We may premise that either case requires more knowledge of surveying 



