INSTRUCTIONS. 5<, 



unexpected discrepancies to bad observation rather than to their much more 

 probable cause. After learning how to make good observations the observer 

 should place the utmost confidence in them, and never yield to the tempta- 

 tion of changing them because they disagree with some preceding observa- 

 tions. Such discrepancies are in general an indication of good, rather than 

 poor, work. 



Stations or tripods which have been unequally heated by the sun or 

 other source of heat usually twist more or less in azimuth. The rate of 

 this twist is often as great as a second of arc per minute of time, and it is 

 generally nearly uniform for intervals of ten to twenty minutes. The effect 

 of twist is to make measured angles too great or too small according as they 

 are observed by turning the microscopes in the direction of increasing gradua- 

 tion or in the opposite direction. This effect is well eliminated, in general, 

 in the mean of two measures, one made by turning the microscopes in the 

 direction of increasing graduation and followed immediately by turning the 

 microscopes in the opposite direction. Such means are called combined 

 measures or combined results, and all results used should be of this kind. 

 As the uniformity in rate of twist can not be depended on for any considera- 

 ble interval, the more rapidly the observations on an angle can be made 

 the better will be the elimination of the twist. The observer should not 

 wait more than two or three minutes after pointing on one signal before poind- 

 ing on the next. If for an)' reason it should be necessary to wait longer, it 

 will be best to make a new reading on the first signal. 



The precision of centering an instrument or signal over the reference or 

 geodetic point increases in importance inversely as the length of the triangu- 

 lation lines. Thus, if it is desired to exclude errors from this source as small 

 as a second, one must know the position of the instrument within one-third 

 of an inch for lines a mile long, or within 6 inches for lines 20 miles long. 

 The following easily remembered relations will serve as a guide to the re- 

 quired precision in any case: 



1 second is equivalent to 0.3 inch at the distance of 1 mile. 



1 second is equivalent to 3.0 inches at the distance of 10 miles. 



1 second is equivalent to 6.0 inches at the distance of 20, miles. 



1 minute is equivalent to 1.5 feet at the distance of 1 mile. 



