MAGNETIC FIELD. 



In the compasses of MM. Brunner, all these reversals and in- 

 versions do not give variations exceeding 5', but we should add 

 that so great a perfection is rarely attained. 



1166. In order to observe the inclination directly, the direction 

 of the magnetic meridian must be known ; it is determined by the 

 property that a dipping needle stands vertical in a plane which is 

 perpendicular to the meridian. 



With a good needle, it is sufficient to take the mean of two 

 azimuths in which the needle is vertical before and after being 

 reversed. It would be sufficient to sight a point in any given 

 azimuth, then to turn the arrangement until this point is in the 

 same position in reference to the vertical circle. One of the bi- 

 sections of these two corresponding positions gives the meridian, 

 aad the other the azimuth perpendicular to the meridian. Except 

 when the inclination exceeds 70, it is easy to find rapidly by this 

 method the direction of the meridian to less than a few minutes, 

 and it is sufficient that the error be less than 1. 



Compasses do not generally enable us to determine directly 

 very small inclinations. Near the equator it is better to make the 

 observation in two rectangular azimuths, or in any two azimuths 

 on either side at the same distance from the meridian. 



In this latter case we need not know exactly the direction 

 of the meridian. The distance of the azimuth planes being 20, 

 for instance, the mean of the observations gives the inclination at 

 10 from the meridian. The .true inclination is obtained by the 

 equation 



tan I = tan I a cos a. 



1167. The measurement of the inclination is always the result 

 of a great number of different readings. It is important to arrange 

 the observations methodically in order to simplify manipulation, and 

 touch the needle as little as possible. One of the faces of the 

 needle has on one side a distinctive mark for instance, a point 

 marked on the mounting near the axis, so that the positions of the 

 magnet can be defined in each case. 



Suppose, for instance, that the observations are made in two 

 rectangular azimuths, nearly 45 on either side of the meridian. 



In order to magnetise the needle, it is placed in a cavity rather 

 deeper than the thickness of the needle and is kept there by a 

 screw. With two magnets the breadth of which is greater than 

 that of the cavity, so that they do not rub against the needle, a 

 number of determinate passes is made 10 or 20, for instance 



