76 A MANUAL OP TOPOGEAPHIC METHODS. 



if possible, be a point similarly well located. The line should, if practica- 

 ble, follow a railroad, in order to obtain the easiest possible grades, and 

 thus avoid errors incident to slope. 



The instrument used for measuring directions should have a circle 6 to 

 8 inches in diameter, and should read by vernier to 10 seconds. The the- 

 odolites formerly used in the primary triangulation are generally used in 

 this work. A larger or more elaborate instrument is not advisable on 

 account of the difficulties of transporting it and frequently setting it up. 

 Upon short lines instruments reading to minutes may be used. 



The readings should be upon signals consisting of poles, and fore and 

 back rodmen must be employed for carrying and setting them. The 

 angular measurements between the poles should be read by both verniers, 

 and it is advisable to note the compass readings at the same time, in order 

 to avoid gross errors. At intervals of 10 to 20 miles, depending upon the 

 number of courses to a mile, observations should lie made for azimuth, 

 observing for this purpose upon the pole star, preferably at elongation. 



The measurements of distance are effected by the use of steel tapes, 

 and preferably by 300-feet tapes, similar to those used in measuring base 

 lines. Two chaimnen should be employed, and in order to avoid errors in 

 the count, it is well to count the rails, in case the work is done upon rail- 

 road tracks. 



The temperature should be noted by means of thermometers at frequent 

 intervals, in order that the proper corrections may be applied. 



The errors incident to running primary traverses are of two classes: 

 errors of direction and errors of distance. 



Those of direction are similar to those treated of under the head of 

 Instructions for the Measurement of Horizontal Angles, and need not be 

 specified here. 



( >\ving to the necessity of setting up the theodolite at frequent inter- 

 vals, it is impracticable to observe at each station the series of angles speci- 

 fied in the above-mentioned instructions, and only a single or at the most 

 a double measure of the included angle, with the reading of each vernier, 

 is practicable for the measurement of direction. It is here provided that 

 observations for azimuth upon Polaris should be much more frequent than 

 in triangulation, and thus an absolute correction to the directions is intro- 



