January 22, 1920] 



NATURE 



549 



Condensers for the proper well-regulated illumination 

 of microscopic objects are identical in optical design 

 with objectives, the only difference being that the light 

 passes through in the reverse direction, and that a 

 lower degree of correction is sufficient not only on 

 theoretical, but also on practical grounds, for nearly 

 always condensers are used in conjunction with the 

 "plane" mirror, which invariably is very far from 

 optical perfection, and so introduces irregular aberra- 

 tions of unknown magnitude and kind, and, moreover, 

 the light from the condenser has to pass through the 

 slide on which the object is placed. This slide is prac- 

 tically little better than window-glass so far as optical 

 quality and perfection of surfaces are concerned, and 

 the great variation in thickness is another source of 

 imf>erfection, especially with dry condensers of high 

 N.A. 



Moderate amounts of residual aberrations in con- 

 densers can always be effectively neutralised by using 

 a sufficiently lar^e source of light of uniform bright- 

 ness or by magnifying the source by a sufficiently well- 

 corrected' "bull's-eye," if the source of light is 

 naturally small. 



A great and very serious defect in the construction 

 of nearlv all condensers of the present day, with the 

 exception of the modest Abbe condenser of two simple 

 uncorrected lenses, is that the iris and the ring for 

 dark-ground stops are placed too far from the back 

 lens instead of being close to the anterior focal plane 

 of the condenser. It is easily shown that such a 

 remote iris-opening or dark-ground stop produces 

 decidedlv oblique illumination of the extra-axial points 

 of the object. With direct light this leads to an un- 

 desirable variation in the type of image and in resolv- 

 ing power in different parts of the field. With dark- 

 ground illumination the result is even more serious, 

 for it is then necessary to use a far larger central stop 

 to secure a dark background over the whole field than 

 would suffice if the stop were placed close to the 

 anterior focal plane of the condenser; such an un- 

 necessarily large stop is highly objectionable, because 

 it reduces the visibility of the coarser structures in the 

 object. 



The increasingly bad position of the ins m the con- 

 densers of hicher' power and shorter focal length sup- 

 plies oracticallv the whole explanation of the universal 

 experience that hijrh-power condensers will not work 

 satisfactorily with low-power objectives, especially for 

 dark-ground illumination. 



The great thickness of the mechanical stage in 

 English stands of the highest quality is the chief 

 reason why the iris and " turn-out ring" of high-power 

 condensers have to be mounted so far below the back 

 lens, and a profound modification of the design^ of 

 the stage with-the view of makincf the part projecting 

 over the condenser as thin as possible therefore appears 

 to be the most desirable improvement of microscope 

 stands from the optical designer's point of view. 



.^s regards the actual making of microscojye objec- 

 tives, it must be borne in mind that the excellence of 

 a computed lens system may be completely swamped 

 by comparatively slight imperfections of workmansTiip. 

 and that high accuracy in this respect is therefore of 

 the utmost importance'. In lenses of high N.A. com- 

 putation shows that a departure from the presrriberl 

 radii and thicknesses by a fraction of a thousandth of 

 an inch mav lead to a' notable loss of perfection, and 

 the polished surfaces must also be truly spherical 

 within less than half a wave-length of light. These 

 limits can be easily observed if modern rnethods of 

 iT-auging and measuring are adopted, and if all sur- 

 faces are polished to accurately made and con- 

 scientiously used test-plates. The tools and methods 

 employed in -really manufacturing lenses on this 

 system were shown by Messrs. W. Watson and Sons, 

 NO. 2621, VOL. 104] 



I Ltd., at the exhibition at King's College in January, 

 1917, and will be found described and illustrated in 

 the record of that exhibition. 



In old English practice the component lenses of 

 microscope objectives and condensers used to be fixed 

 in their cells by cement of the sealing-wax type. 

 Many old lenses which are still found in perfect ad- 

 justment fifty or more years after being mounted 

 I demonstrate that the cement may hold the lenses in 

 correct position almost indefinitely ; but other experi- 

 ences, especially with lenses used in tropical countries, 

 suggest that shifting may occur, and it is therefore 

 strongly to be urged that all microscope lenses should 

 be held between metallic shoulders at both ends by 

 being bevelled into their cells, care being naturally 

 required to avoid pressure and distortion through too 

 tight a fit. 



A point on which users of objectives err to their 

 own detriment is an excess of faith in numerical aper- 

 ture. I have heard microscopists boast of possessing 

 an objective, say, of 1-43 N.xA., whereas somebody else 

 had one of barely 1-40 ; and a careful test would 

 show that whilst the 1-43 was an indifferent lens, the 

 140 was excellent. The fancied advantage of 2 per 

 cent., then, is really a disadvantage of perhaps 25 per 

 cent, or more. 



One of the few disservices which Abbe did to micro- 

 scopy was the pushing of the N.A. of dry lenses to 

 095, and to a less extent the increase of that of oil 

 lenses to 1-40. The extreme marginal zone of the 

 apochromatic dry objectives of 095 N.A. is particularly 

 badly corrected, so much so that the lenses will only 

 bear a solid illuminating cone of about 065 N.A. even 

 on the Abbe test-plate, and that with annular light 

 ! bringing only the marginal zone into action correction- 

 j collar and tube-length combined do not allow of 

 reaching a point of good spherical correction. There 

 is no doubt that Abbe's own earlier dictum still holds, 

 to the effect that beyond about 0-85 N.A. the higher 

 aberrations become unmanageable unless the free 

 working distance is reduced to a very few thousandths 

 of an inch. A carefully computed objective of 

 085 N.A. will bear a full illuminating cone on suit- 

 able objects, and can thus realise its fullest resolving 

 power. An objective of 095 with a condenser of 0-65 

 has the resolving power of the mean, or of o-8o N.A., 

 and is thus actually inferior, except for freak resolu- 

 tions, with extremely oblique light. Oil objectives 

 of more than 130, or at most 1-35, N.A. are also of 

 very doubtful added value. 



In closing this section I will once more quote with- 

 out comment an anecdote of Fraunhofer, who received 

 a complaint that a telescope supplied by him, although 

 giving magnificent images, displayed certain fine 

 scratches when examined with a magnifying-glass ! 

 The reply sent by Fraunhofer is reported to have 

 been : "We have constructed the telescope to be 

 looked through, not to be looked at." 



A few sentences may perhaps be added as to the 

 prospects for further improvements of microscopic per- 

 formances. I have stated earlier in this paper 

 that there is a bright ray of hope with regard to 

 diminishing the curvature of field without loss of 

 definition. 



Advances in numerical aperture offer very little 

 attraction. Abbe, in my opinion, carried the N.A. too 

 far rather than not far enough, and I am not aware 

 that any notable discovery has been achieved with the 

 few monobromide immersion objectives of N.A. 160 

 which he designed. 



The use of a shorter wave-length, i.e. ultra-violet 

 light, is a little more promising. There would 

 be none but technical difficulties to the construction 

 of lenses suitable for this work. But as only very few 



