124 



Messrs. Thorpe and Riicker on some 



packed in the ordinary magnetometer-box, and when it is 

 in position, and the platinum mirror is directed towards the 

 sky, or, if desirable, towards its brightest part, the image of 

 the wires can readily be seen. If the background reflects 

 much light, it is desirable to place a small blackened metal 

 screen weighted with lead on the horizontal circle which 

 carries the mirror. This can be lifted off without in any 

 way interfering with the adjustments. 



The following two examples may serve to indicate the 

 degree of precision with which the adjustment of the mirror 

 can be made. 



Observations were taken both when the observer was look- 

 ing towards and directly away from the sun. The N point 

 on the circle is deduced from the front observations. The 

 collimation correction is obtained by combining the front and 

 back observations. It will be seen that the uncorrected 

 results are quite as good as those to which the collimation 

 correction has been applied, which proves that the mirror 

 can be adjusted to within the limits of the error of observation 

 on the sun. 



Station. 



G. M. T. 



Altitude 

 of Sun. 



N point on 

 Circle. 



Collimation 

 correction. 



Corrected 

 N point. 



Horsham. ( 



(April 21, \ 



1888.) ^ 



h. m. s. 

 12 43 49 



15 15 44 



17 4 



5°0 2 ! 2 

 34 33-7 

 18 12-7 



297 4-2 

 297 3-8 

 297 3-6 



+0-3 



-0-6 



01 



297 4 : 5 

 297 3-2 

 297 3-5 



Alresford. f 



(April 28, \ 



1888.) [ 



14 10 36 



15 45 6 



45 7-8 



241 23-0 

 241 22-5 



-0-1 

 Not det 



241 229 









In cases where it is for any reason impossible to take the back 

 observations, or if the sun is so frequently covered that it is 

 advisable to observe without waiting to adjust the mirror, it 

 is advantageous to be able to determine how much error has been 

 introduced. For this purpose we have placed in the plane of 

 the cross wires a fine scale engraved on ivory. 

 The angular deviation of the mirror which 

 produces a given displacement of the image 

 of the cross wires on this scale can be deter- 

 mined, and when this is known the error of 

 the mirror can, on any subsequent occasion, 

 be deduced from the deviation of the image. 



