SEXTANTS AND OTHER REFLECTING INSTRUMENTS. 65 



the other, an aiTangement is made for placing the shades to be tested in such a 

 way that they can be turned round. Having placed in this fi'ame a shade for 

 trial, I bring the two images of the sun seen in the telescope into contact, by 

 means of the micrometer screws ; and if, on turning the shade before the object- 

 glass, the contact remains perfect, then we know that its surfaces are parallel ; if, 

 on the other hand, the contact is broken, we can, by turning the shade, find the 

 point of nearest contact; and by turning the shade through 180°, we measure the 

 amount of opening of the images, by means of the divided micrometer head. This 

 gives double the error which would be caused by the shade when used in making 

 a direct observation. But as all angles observed by reflecting instruments are 

 double, or, in other words, the divisions on the Umbs of reflecting instruments are 

 equal only to half those of an instrument used for direct observation, the double 

 error given by this method goes all to deteriorate the observation, as an unparallel 

 shade affects the angle when reflected by twice the amount "of direct observation. 

 Another important object is gained by this method of testing the shades. It being 

 a very difficult matter to obtain these altogether perfect, I reject aU where the 

 error would amount to lO'', which is the usual reading of a sextant ; and when an 

 error of a smaller amount does exist, I find, by turning the shade, the line in 

 which the want of parallelism exists ; and by placing this line at right angles to 

 the plane of observation in a reflecting instrument, the angles observed by such 

 an instrument are not afffected even by the small error of the shade. 



VOL. XVI. PART I. 



