TERRESTRIAL MAGNETISM 307 



error due to the component of the displacement of G perpendicular 

 to the axis is eliminated on taking the mean of the readings before 

 and after turning over. The face of 

 the instrument is then turned round 

 into its original position facing East, 

 and the positions of the ends of the 

 needle are again read. Denoting 

 the aspects of the face of the in- 

 strument by E and W and those of 

 the face of the needle by e and a>, 

 we have to read each end in each of FIG. 229. 



the four arrangements Ee, Wo>, We, 



Ew, and the mean of the eight readings taken eliminates the errors 

 so far considered. 



There is still outstanding the error due to OB, Fig. 229, the 

 component of the displacement of the centre of gravity parallel 

 to the axis. To eliminate this the needle is taken out and its 

 magnetisation is reversed by the method of divided touch. It is 

 assumed that the magnetisation is exactly reversed an assumption 

 quite certainly not fulfilled, but it at least gives us a method of pro- 

 cedure which tends to diminish the error. B is now thrown as much 

 above O as it was previously below it, and going through the four 

 cases again the mean of the readings with the magnetisation direct 

 and reversed gives as good a value of the dip as can be obtained with 

 a single needle. It is usual to find the value with two different 

 needles and to take the mean in the hope that it is nearer the true 

 value than that given by either separately. In this country, with 

 a value near 70, skilled observers may obtain determinations with 

 different needles differing by one or two minutes, and different 

 instruments may give results differing by quantities of the same 

 order. 



We assumed that the observations began with the circle in the 

 magnetic meridian. To make this adjustment it is usual to set 

 the microscope verniers at 90, and turn the circle facing S till the 

 ends of the needle are on the cross wires. The position on the 

 horizontal circle is then read. The needle is then turned round and 

 the dip circle is turned round the vertical axis till the microscopes 

 again sight the needle-ends on the cross wires, and another reading 

 on the horizontal circle is taken. The dip circle is then turned 

 through 180 to face North, and the operations are repeated. The 

 mean of the readings on the horizontal circle then gives the 

 position of the circle in which the magnetic axis of the needle is 

 vertical and, there fore, that in which the plane of the circle is perpen- 

 dicular to the magnetic meridian, for in that plane the vertical force 

 alone can move the needle. The dip circle is then turned through 

 90 and the needle is assumed to be in the magnetic meridian. 



Sometimes the dip is taken by another method which does not 

 require this adjustment. 



