THE USE OF THE MAGNETIC NEEDLE IN SURVEYING. 1 63 



is the case, by directing the sights to various objects in different direc- 

 tions, reversing on each, and noting whether the needle gives a different 

 reading by any of the reversals. If the reading is not changed by re- 

 versing in any of the positions, the box may be considered free from 

 magnetic power. 



4. The pivot should be exactly in the centre of the graduated circle, that 

 the two ends of the needle may give the same reading. It is true this 

 error may be eliminated by reading both ends, and taking the mean ; but 

 it is better to have no error to eliminate. If, however, the two ends do 

 not read alike, the mean reading should be used. 



5. In ttsing the compass^ proximity to all magnetic substances, both 

 natural and artificial, must be avoided. Masses of iron, like gas- or 

 water-pipes, water-conductors, and lightning-rods, are a source of disturb- 

 ance not easily avoided in cities ; and the water-conductors and lightnings 

 rods being placed vertically are more disturbing than larger masses lying 

 horizontally. The reason of this is that, having their longer axis more 

 nearly in the direction of the force of terrestrial magnetism, they become 

 magnetic by induction, and act not merely as so much iron, but as mag- 

 nets. They must therefore be given a wide berth by the surveyor with 

 the magnetic needle. In making observations with the magnetometer, it 

 is thought necessary to remove from such objects to a distance at least 

 equal to twice their height ; but it is probable that no perceptible influ- 

 ence upon an ordinary needle would be observed at half that distance. 



Besides these larger masses of iron, the surveyor sometimes carries upon 

 his person the cause of much error. Ordinary knives, if not brought 

 nearer to the needle than two or three feet, will have no appreciable 

 effect, but magnetized knives should be kept at a greater distance ; and 

 the chain-men should not be allowed to bring the chain within less than 

 twenty feet of the instrument. There is another source of disturbance, 

 carried by the surveyor himself, which frequently he does not suspect. It 

 is the common buttons, with an iron body, used upon coats. In reading 

 the bearing, these are likely to be brought near to the needle, and to pro- 

 duce considerable deviation. Such buttons ought not to be worn in 

 working with a compass. 



Still more difficult is it to avoid local attraction by magnetic rocks, 

 which are more common than is generally supposed. Indeed, so common 



