114 
HINTS TO TRAVELLERS. 
to a distance of three miles with a somewhat greater error; and an 
adaptation of the process is capable of yielding reconnaissance traverses 
and approximate trigonometrical work far more accurately and expedi¬ 
tiously than can be looked for by any other means, unless a regular 
trigonometrical survey be resorted to. 
The theodolite used should be six-inch or larger; it should be simple 
in construction, and furnished with one vertical and one horizontal wire. 
The bars may be of varying lengths. In the Himalayas the 20-foot bar 
was in general use, but ten and two-foot bars were found convenient for 
some purposes. A 20-foot bar with 12-inch circular discs * is capable of 
furnishing, under favourable conditions of light and atmosphere and by 
a skilled observer, a 3-mile distance with an error of six feet. A ten- 
foot bar with eight-inch discs will give good results up to a mile and a 
half, and a two-foot Gunter’s scale blackened at the ends with two-inch 
paper discs pasted on two feet apart, and properly mounted, will give 
distances up to 20 chains. 
The modus operandi of a traverse surveyor must now be explained in 
detail. 
The forward signalman sets up the horizontal bar over the station 
mark, and then, by means of the folding sight-vane, directs the bar at 
right angles to the observer, who then intersects and records the reading 
of the back signal. Then, leaving the lower clamp fast, he releases the 
upper plate and intersects the right-hand disc of bar, the reading of 
which he records. 
Now release lower clamp (leaving upper clamp fast) and intersect left- 
hand disc. Again release upper plate and intersect right-hand disc, and 
for a second time the left-hand disc with lower plate, and so on, con¬ 
tinuing the repetition until, say, ten times, and then read and record the 
right-hand disc. In this operation the graduated limb of the theodolite 
will have moved over an arc ten times greater than that subtended by 
the bar. Now repeat again, ten or twenty times, and record readings of 
right-hand disc, and then, having taken a vertical angle to bar, and 
leaving lower plate fast, intersect, and record the reading of back signal 
with upper tangent screw, and such a record as I here show will have 
been obtained:— 
* For illustration, see p. 38. 
