Chap. 7] GRAVITATIONAL METHODS 239 



in the form of eq. 7-85. The general form of these integrals is 



If-/. "x da, which rauy be evaluated by the integraph as surface 

 J J p (sm na) 



integral of the form ± - , fiv A more detailed discussion 



of this integraph is given in the following section on interpretation of 

 torsion balance anomalies. 



(d) Field practice in terrain survey; preparation of station site. The 

 accuracy of a torsion balance may be enhanced considerably by a judicious 

 selection of the station site. The location should be as flat as possible, 

 at least to a distance of about 20 meters. Vicinity of ditches, bluffs, 

 embankments, and houses should be avoided if possible. In forests it is 

 advisable to set up the instrument in such a manner that large trees are 

 symmetrically arranged with respect to it. If setups in rugged terrain 

 have to be made, the site should be so selected that the effect on gradients 

 is a minimum. The site should be leveled with a shovel to a radius of 

 3 to 5 m. In the filling of depressions, the dirt should be tightly tamped 

 to preserve approximately the same density around the station. Leveling 

 the immediate vicinity has the further advantage of avoiding difficulties 

 in setting up the house. 



After the site has been leveled, a plane table or transit is set up and 

 the "far terrain" determined first (at 5, 10, 20, 30 m, and so on). On the 

 plane table sheet, eight or sixteen azimuths are laid off. If a transit is 

 used, it is advisable to mount a compass on top and to read the angles on 

 it instead of on the horizontal circle. The instrument is set up over the 

 station point. One man reads the elevations and takes them down, the 

 other goes around the station and places the leveling rod where directed. 

 A rope or chain with the proper measurements marked on it is used in 

 obtaining distances. It is advisable to carry a rope 20 m long for the 

 average terrain and an extension out to 100 m for more complicated 

 terrains. The rope is laid out first in the north direction and the rod is 

 placed at 5, 10, 20, 30 m, and so on. Then the rope is moved over 

 to the next azimuth and the rod is placed at 30, 20, 10, 5 m, and so on. 



In some terrains the use of ropes or chains may be impracticable. Some 

 companies have trained their men to pace the distance accurately. In 

 brush country it may be necessary to cut lines for the terrain survey 

 (Fig. 7-87). The rod should be graduated in centimeters; it may be so 

 made that differences in elevation may be read directly. For this purpose 

 the rod is graduated both ways from zero, with red divisions below and 

 black divisions above, red indicating positive elevations and black negative. 

 The foot of the rod carries an adjustable shoe. Before elevations are 

 taken, the rod is placed close to the telescope of the transit and so ad- 

 justed that the zero point coincides with the optical axis of the telescope. 



