302 On a iiexc Scxfanl recentlij iiwcntcd 



total reflection will be aoain reflected upon the side A C, and 

 go out parallel to their hrst direction : all these rays passing 

 through the object-glass will form at its focus the image re- 

 flected from the point Q, which will fall exactly upon the di- 

 rect image formed in this same focus by parallel rays which 

 come from the imcovered half of the object-glass, by means 

 of which the point zero of the graduated limb may be ascer- 

 tained. Nevertheless if we turn the prism round its edge A, 

 on the side B C A, it will present new objects in succession 

 coinciding with the object Q, until the side AC shall be parallel 

 to the object-glass ; we shall then have the super-position of 

 all those points that are 90 degrees distant from the point Q. 

 It is therefore clear, that by this means we can measure all 

 angular distances as far as 90 deg. and a little more : I say a 

 little JiiOfe, because that depends on the extent of the refraction 

 of die glass* ; and if the prism be still turned on the same side, 

 the total reflectioii of the rays upon the plane BA will no 

 longer take place. 



If^ before the other half of the object-glass, which we have 

 supposed not to be covered, a second prism, equal to the first, be 

 placed, but moveable in a contrary direction, the two objects will 

 then be equally seen by reflection ; and by the combined move- 

 ment of the two prisms we can carry the measure of an angle to 

 the double of the greatest angle measured by one prism alone. 



These principles, upon which rest all the theory of angular 

 measures by these prisms, being well understood, it will be 

 easy to comprehend the construction and the use of this new 

 instrument. Fig. 2 shows it in perspective, as it was ti'aced 

 in a camera lucida. 



A B D is a sector greater than a quarter of a circle of four 

 inches radius. The limb is divided fi-om 10 to 10 minutes, and 

 by means of the vernier we can read to 10 seconds. About 

 the centre C turns the index CE, which carries the vernier 

 at one extremity, and the isosceles rectangular prism F at 

 the other, with its edge 3 C directed towards the centre and 

 pei'pendicular to the plane of the limb. The other prism H, 

 equal to the first, is fixed in its place on the instrument, 

 and so disposed that when the index marks zero, the larger 

 faces of two prisms are nearly in contact and jierfectly parallel. 

 Lastly; a telescope N, supported by the arm LI, is moveable 

 on the plane of the sector about the centre C. M is a micro- 

 scope by which to read the divisions : and this forms the whole 

 of the instrument. 



The eftect of this double motion of the telescope parallel to 



* Witli a pribin of common glass anglts as far as 102 degrees may be 

 niciisiirecl. 



the 



