GANNETT.] METHODS OF GEOGRAPHICAL WORK. 679 



ments of stone, 5 feet in lieiglit, were erected at the time of occupation. 

 Previous to occupation the natural summit of the station was sighted. 



The topography was secured as usual by maj) and perspective sketches. 

 The former form the basis of the map. They were made by eye on an 

 assumed scale, the distances and directions being estimated. The as- 

 sumed scale of these sketches was one mile to an inch. In transferring 

 these sketches to the map in the office, their inaccuracies are corrected 

 by the exact location of all important points, as mountain peaks, buttes, 

 angles of plateau, mouths and bends of streams, &c. In the area sur- 

 veyed this season (13,000 square miles), about 1,000 points, besides the 

 350 stations and substations, were located by intersections of sight- 

 lines, making, with the stations, a located point in every 10 square 

 miles. 



Elevations were measured by the mercurial barometer, aneroids, and 

 the vertical arc of the gradienter. During the season, base barometric 

 observations were taken at the adjutant's office at Fort Hall, Idaho. 

 The height of this point was determined by the computation of coincident 

 barometric observations for four months at this point and Oorinne, Utah. 

 Through the kindness of the Chief Bignal Officer of the Army, the 

 barometric observations at Salt Lake City also were furnished me for 

 use as base observations. 



The heights of the camps, where more or less extended series of ob- 

 servations were taken, were determined by direct reference to the base 

 stations. The camps ranged in height from 4,500 to 7,000 feet, only one 

 or two exceeding the latter height ; h&nce in these cases there was lit- 

 tle room for error due to the defects in the barometric formula. The 

 heights of stations were determined by barometric reference to the 

 camps at their bases, and by vertical angles with the gradienter from 

 them, and, in addition, the stations were connected with one another by 

 a complex system of vertical angles, in such a way that their relative 

 heights were known, and thus all barometric measurements were reduced 

 to a common point, so that the height of each station was a mean of all 

 barometric measurements of stations. The heights of all located points 

 were measured by dip angles. The heights of points of minor impor- 

 tance were measured by aneroids, and referred for computation to the 

 nearest camp. Aneroids were compared, every morning and evening, 

 with the mercurial barometer. 



From the perspective sketches, aided by these measured heights 

 (which number in the aggregate more than 1,500), the distribution on 

 the map, of contours 200 feet apart vertically, has been effected with a 

 considerable approach to accuracy. 



