34 A SYSTEM OF ACCURATE MEASUREMENT 
The precision with which the theodolite observation of the 
altitude of the chord of the riband must be observed will depend 
upon the requirements of the case. ‘These will be limited by . 
rapidity with which the cosine of the angle changes, and b 
accuracy of the instrument available. A properly-fitted 5” fren 
dolite may be depended upon to give angles within 15” of their 
absolute value. 
Some general remarks in regard to the details of this system of 
measurement will not here be out of place. 
The friction of a Ticht riband, when under a moderate tension 
upon ordinary ground may, in general practice, be neglected without 
appreciable error. When the surface is slightly uneven and the 
riband lying in a succession of little curves, a small inerease of 
tension, the amount of which may be determined by a few simple 
experiments, will sufficiently correct the consequent defect. 
Chaining upon the aaa is most convenient when it closely 
approximates a plane surfac 
_ The terminal marks ‘eal the riband should be two or three 
inches short of the ends, so as to permit their easy and accurate 
adjustment to any given points. 
The spring balance should be as close as possible to one of these 
marks, or may be so attached that its index-point will define the 
end of the chain. he latter method has the advantage of indi- 
cating the register of the tension, and the a to mark at 
together, so that they may be observed simultaneously. 
teel ribands marked at hundredth links are now obtainable, 
and by means of them the fractions of the chain may be recorded 
to almost any assigned practical degree of accuracy. 
For facility in Sah ae computation, whole chains should 
laid out when pos 
The sane in the Ficld-book may be conveniently made thus:— 
50000 links D 20°12'15” F 72/30=469°357 
22°105 E_ 1° 30’ F 74/30= 22-103 
ifies depression, E elevation, F Fahrenheit, the subscribed 
30 the temperature of the riband when at standard length. 
Tn conclusion, it may be observed that the time occupied by the 
adoption of the system in practice is inconsiderable. If the strict 
ment yet adopted, and under all circumstances 
gives results diduadiin g in accuracy any of the methods preceding it. 
the Appendices will be found a mathematical discussion of the 
whole problem, and also tables for computing the Se for any 
ri ith a considerable variation of standard te 
= Nore —1 desire to express my indebtedness to Mr Mr. F. B. W. 
Tigi District beh and to H, 8. Hawkins, MA., of the 
of this subject. 
