in the lug of l ( 'ig. 69, that can be tightened from time to time as the 

 ion requires. Beyond a comparatively short interval each side 

 of the horizon, however, the gradienter ceases to be accurate be- 

 cause the pitch of the screw gradually looses its strict proportion 

 with the chord of the arc it is supposed to measure. 



The discussion on the gradienter, pp. 1 5 and 24, does not profess 

 to take vertical angles into consideration while working close to 

 the hori/on. 



Hypsometric Leveling 



This practice with the gradienter consists of determining a 

 vertical angle in terms of its tangent and having measured the in- 

 clined distance with the tape, or with the gradienter itself, to find 

 the difference in elevation by multiplying the measured distance by 

 the sine of the angle found. 



For instance, suppose the gradienter to be turned 3 revolutions 

 and 18 spaces below the hori/on in order to strike a certain station 

 peg. This movement is equal to 168 spaces as previously explained 

 or a gradient of 1.68 %. Referring to table, p. XX VI, we find this! 

 equal to an angle of 57' 45", the natuial sine of which can be used 

 to get the difference in elevation. 



Tables compiled from the Army Manual may also be used for 

 this purpose. Having found the percentage equivalent in degrees 

 simply add the multiple parts as when using an ordinary traverse 

 table. (See p. XXVII Apx.) 



Gradients 



Let it be required to run on a gradient of 1% compensated to 

 1.94%. Turn the drum three complete revolutions and 44 additiona 

 spaces either in elevation or depression. Take the H. I. and ob 

 serve the rod at the several successive stations. If H. I. = 5, am 

 the rod reading is 4, a cut of 1 ft. would be implied. If the roc 

 reading is 6, a fill of 1 ft. w ill be implied, etc. This is the mos 

 facile method known for running grades in highway or railroac 

 construction. (See also page 24.) 



To Measure Horizontal Deflections with the Gradienter 



Contributed by Vrof. D. (.'. Humphreys, Washington and 

 Lee Univ., Lexington, Va. 



When the compass was good enough for land surveys, a tree 

 that happened to intercept an alignment \\as good enough fora 

 red. apd the survey pron-eik-d In s'-ttinu up on the other side anc 

 continuing the line on the same magnetic boaiing. Now, however, 



124 



