February 25, 1909] 



NA TURE 



4S9 



VLTRA-MICROSCOPIC VISION. 



IN Nature, November 5, 190S, a short paragraph 

 appeared in reference to a letter received from 

 Mr. G. V. Raman, of the Science Association Labora- 

 tory, Calcutta, referring to a method of dark ground 

 illumination for the microscope. From it, and from 

 •a subsequent communication on the same subject, it 

 uould appear that the subject of dark ground illumin- 

 ation and ultra-microscopic illumination may in 

 certain directions give rise to controversy, and result 

 in some confusion of thought. 



It is unquestioned that any method of microscopic 

 illumination in which the direct axial beam of light 

 is cut out, and where, therefore, a grazing or oblique 

 illumination is obtained, may result in making 

 visible some particles that are beyond the limits of 

 ;i microscope illuminated by ordinary methods. 



It must at once be admitted that it is difficult tu 

 define the exact boundary beyond which objects may 

 be said to be ultra-microscopic. To appreciate this 

 ix)int, it is necessary to refer very briefly — owing to 

 the limits imposed in such a short article as this — 

 to the wide difference between the limits of micro- 

 scopic resolution and microscopic visibility. 



To define the limits of resolution of the microscope 

 is not diflicult, as this is purely a function of the 

 numerical aperture of the objective. The limits in 

 this direction have been accurately determined, and 

 practically agree in theory and practice. In the case 

 of periodic structure, such as in diatoms, or in 

 mechanically-ruled plates such as Grayson's rulings, 

 this resolving limit can be found by multiplying the 

 numerical aperture of the objective by So, 000 when 

 monochromatic green light is used, and illumination 

 is by a solid axial cone of light. This means approxi- 

 matelv that lines of more than 120,000 to the inch 

 would be beyond the limit of resolution when using 

 an objective with N..^^. r40, the largest aperture 

 generally available at present ; or that two points 

 lying closer together than the distance between these 

 would be evident, not as two separate images, but 

 would so overlap as to appear as one. 



This, however, is by no means the limit of visi- 

 bility, and Lord Rayleigh states that isolated objects, 

 or two bright areas separated by a dark line, may 

 be seen if the dark line is as narrow as i/i6, and 

 under certain conditions 1/32, of a wave-length of 

 light, although the resulting image does not of neces- 

 sitv represent the actual appearance of the object. 

 The flagellum of a bacterium, for instance, may be 

 much bevond the limit of resolution, but is visible 

 because it is an isolated object. 



Another factor is the intensity of the incident light, 

 and there is some reason to conclude that any rela- 

 tively isolated object may be visible if it is illuminated 

 with' sufificient intensity, and can reflect light enough 

 for the eye to appreciate. A keen observer will see 

 in a micrcscopic image all structure that the best 

 objectives can reproduce with a magnification of little 

 more than 750 diameters, although it may be con- 

 venient to amplify the image beyond this to facilitate 

 observation. Objects that are smaller than this limit 

 of resolution are generally referred to as ultra- 

 microscopic, although it is obvious that the term is 

 not always jusiifiod. It is clear, therefore, that to 

 define the moaning of the term " ultra-microscopic " 

 is by no moans simple, and especially in view of the 

 fact that most methods of dark ground illumination 

 do result in the formation of images that are not 

 seen in other ways. 



Illumination in the microscope by means of light 

 projected at various angles to the optical axis has 

 been common for very many years. The writer has 

 used, for example, an oil immersion paraboloid made 



NO. 2052, VOL. 79] 



by Messrs. Swilt and Son, probably very soon after 

 tlie introduction of oil-immersion objectives, say 

 about 1S75, snd 'he results to be obtained with it 

 compare favourably with those of more recent intro- 

 ductions. At various times, other methods have been 

 introduced. The simplest, and one of the earliest, was 

 what is known as a " spot lens," also a dry para- 

 boloid, and the arrangement by which the ordinary 

 substage condenser may be utilised. In each of these 

 a blackened stop of suitable size is placed beneath 

 the optical portion of the illuminating system in such 

 a position that the central axial rays are obstructed, 

 and no light directly enters the objective. Only light 

 refracted or reflected by the object reaches the objec- 

 tive, and the former, therefore, shows up brightly 

 illuminated on a more or less dark background. With 

 each of these arrangements only objectives of rela- 

 tively low aperture can be used. 



Other methods are those in which a stop is placed 

 above the posterior combination of the objective, or 

 the very ingenious arrangement suggested by Mr. 

 J. W. Gordon, in which the stop, in this case a small 

 globule of mercury, is placed above the eye-piece in 

 the position occupied by the Ramsden disc. The 

 closest approximation to the modern ultra-microscope 

 of Siedentopf is the type of dark ground illumination 

 in which the light is reflected so that it impinges upon 

 the object at right angles to the optical axis of the 

 microscope, but in none of these is any attempt made 

 to confine the illuminating beam to the area under 

 observation. 



In 1903 an entirely new method of rendering visible 

 ultra-microscopic particles was brought out by Sieden- 

 topf, and arose out of some investigations being made 

 on various shades of ruby glass. As is probably well 

 known, the colouring of ruby glass is dependent on 

 small particles of gold, the dimensions of which 

 approach in size to that of a molecule. If examined 

 under an ordinary microscope and by ordinary methods 

 of illumination, or by any method of dark ground 

 illumination, even with the very best objectives there 

 is no indication of the presence of any isolated 

 particles. But, by a method of projecting a very 

 thin cone of light at right angles to the optical axis 

 of the microscope, and exactly on the spot under 

 observation only, they were able clearly to observe 

 diffraction discs which became visible, and arose from 

 each individual particle of gold in the ruby .glass.' 

 The method is therefore one entirely depending on 

 the arrangement and exact control of the illumin- 

 ation. The initial intensity of the illuminant must be 

 high, so that only the electric arc or sunlight is 

 suitable. 



In general, the illumination of the object is accom- 

 plished by projecting the image of a very narrow 

 precision slit, which is constructed in a similar 

 manner to those used on fine spectroscopes so that 

 both the length and breadth tif the slit can be 

 varied and exactly determined on to the object. The 

 apparatus is arranged so that a very powerful beam 

 of light is projected through the slit and focussed 

 by a suitable optical arrangement so that the apex 

 of the cone of rays falls exactly on the spot in the 

 object where the objective is locussed, and so that 

 no particles lying above or below this spot are 

 illuminated. Consequently, only the particles in the 

 field of view are sufficiently luminous to form an 

 image in the microscope, and no particles lying 

 outside this field can diffuse light and interfere with 

 the formation of the image. 



This latter disadvantage is common to all other 

 methods of dark ground illumination, and it is in 

 this respect chiefly, if not entirely, that the Siedentopf 



1 See Journal of the Roj-al Microscopical Society, 1903, p. 575. 



