ADJOURNED DISCUSSION IN LONDON 233 



Discussion. 



Commander M. A. Ainslie, R.N.: With regard to design, the 

 principle of the optical bench seems to me exactly the principle needed 

 in order that you may build up in bits the apparatus you want 

 for any particular research, so that everything may fall naturally 

 into alignment. Each piece of apparatus should be on a separate 

 saddle of its own. I would even have the eye-piece on a separate 

 saddle, with a separate coarse adjustment of its own; this may sound 

 revolutionary, but I believe it to be perfectly sound. Then, again, 

 I think we ought to have a longer range to the draw-tube ; as a 

 rule, it is quite insufhcient, especially when high power dry objec- 

 tives are in use. An ordinary dry 3 mm. objective requires a change 

 of about 20 mm. in the tube-length to compensate for a change of 

 .01 mm. in the thickness of the cover glass; and although objec- 

 tives of lower power are less sensitive, objectives of low power and 

 large aperture are not very easy to obtain. 



With regard to the size of illuminant required in photomicro- 

 graphy, whether of metals or of other objects, this is settled by a 

 very simple relation. If d be the diameter of the light-source, and 

 B that of the illuminated area on the object slide, and if B be the 



^° 



angle made with the axis of the extreme ray entering the optical 

 system and that of the extreme ray falling on the object, the latter 

 being supposed in a medium of refractive index //, then we always 

 have 



d sin = /J- D sin 0, 

 which is, of course, merely the well-known '' optical sine law " ; it 

 really amounts to saying that the product of the diameter of the 

 light-source into the N.A. of the collecting lens is equal to the dia- 

 meter of the circle of illumination on the object, multiplied by the 

 N.A. of the condenser. You cannot get away from this relation; 

 it settles once for all the diameter of the illuminated field, and it is 

 true for any optical system whatever between the light-source and 

 the condenser. 



If you are going to use a metal filament lamp, you are con- 

 fronted with one of two things; either you are going to project 

 an image of the filament on your object, or else you are going to 

 project this image into the plane of the objective aperture, filling 

 it irregularly; a state of things which Professor Conrady long ago 

 show^ed to be incorrect. The diameter of the filament is far too 

 small, having regard to the relation I mentioned just now; and 

 of course one does not want an imago of the filament on the photo- 

 graph. 



With regard to the intensity of the arc, what decides the ex- 

 posure is the intrinsic brilliancy and not the total power of the 

 arc. As to the heating effect, I have used a 25 ampere arc within 



