Mr. G. Forbes. 



[Feb. 22, 



the other half-base, using the whole base with the binocular. Having 

 now got the square fields in the middles of the round fields, and 

 square with each other, these adjustments seldom if ever require 

 attention. All subsequent adjustments are made by the screws on the 

 outer prisms; the object of these adjustments is to make the prism 

 .axes parallel. If the eyes seeing the balloons as one balloon see dis- 

 tant objects double, one image over the other, then the screw which 

 tilts the prism about an axis perpendicular to its axis of symmetry 

 will accurately remove this defect, and here also one prism is adjusted 

 with the half -base first, and then the other prism, with the whole base. 

 The tilting of the prism about the axis of symmetry would be cured 

 by the other screw, but the defect is not easily seen by the binocular. 

 It can however be detected by the collimator and removed by using 

 the adjusting screw, which operation will not afreet the previous ad- 

 justments. 



The collimator is a telescope with a spectroscope-slit, near the eye end, 

 mounted at the end of a tube projecting at right angles to the telescope 

 tube and attached to it near the focal plane, and having a right angled 

 total reflection prism in the axis of the telescope and a mirror with a 

 universal joint for illuminating the slit. The binocular being removed, 

 the collimator is pointed to one of the middle prisms. Four images 

 of the slit are then seen, reflected from the two unsilvered surfaces of 

 ■each of the two prisms in the half-base under examination. They 

 are easily identified by their different brightnesses. The 2nd and 

 3rd in brightness are chosen, and the screw worked until these appear- 

 in the same line. Then the faces of the two prisms which are opposed 

 to each other must be parallel so far as a tilt about the direction of 

 vision is concerned. Adjustment about a vertical axis is not neces- 

 sary. The two adjustments now made render the axis of the two 

 prisms perfectly parallel, and the base is adjusted. The slit of the 

 collimator can be varied in width by a divided screw head. One turn 

 of the screw opens the jaws of the slit fully, when the width is three 

 minutes of arc. This gives a ready means of measuring the error. 

 There are other methods of proceeding to the corrections of the outer 

 23risms. We may begin by using the collimator and adjusting either 

 or both of the screws previously used • and afterwards, with the 

 binocular, and using the third screw, we make the ends of the balloon 

 ropes rest on the same distant horizontal line. In actual warfare this 

 method has certain advantages. 



Addendum, received March 1, 1902. 



Beaufort West, Cape Colony. 

 I have made a large number of tests of the instrument, some days 

 giving better results than others. Instead of choosing the days that 



