MICROSCOPES: ALFRED W. PORTER 73 



then the same applies to the emergent light. In either case there 

 must be diminution of resolution. In the case of metallurgical 

 specimens these defects are absent. The light is reflected from 

 almost a mathematical surface. It may easily be, therefore, that 

 for such specimens fullest advantage may be taken of permissible 

 magnifying power, especially where the detail is of a simple char- 

 acter. This is seen, for example, in Figure 23 of the contribution 

 by Sir R. Hadfield and Mr. Elliot, where the line markings of 

 Pearlite are very clearly portrayed. On the other hand, in Figures 

 24 and 25, where there is evidently much fine detail below the 

 resolution limit, it is not clear that the high magnification used is 

 any advantage. Even if this fine detail appeared sharp, it w^ould 

 have no significance. 



In the metallurgical case it must be borne in mind that if the 

 mirror or prism in the vertical illuminator is opaque, it blocks out 

 part of the aperture. The resolution of lines (such as those of 

 pearlite) will be different, according to the azimuth in which they 

 lie. Taking the aperture as semi-circular, the character of the 

 image of a point is a central oval (instead of circle), the minor axis of 

 the oval being parallel to the bounding diameter of the opening, and 

 about half the length of the major axis (Struve, Mem. de I'Acad. 

 des Sc. de St. Petersburg (7), XXX., No. 8 (1882); Bruns, Astr. 

 Nachr., CIV., 1 (1883); Straubel, Inaug. Disert., Jena (1888); 

 Scheiner and Hirayama, Abhand. Gesell., Berlin (1894); P. F. 

 Everitt, R. Soc. Proc, A83, 302 (1910). Scheiner gives a photograph 

 of the diffraction figure. Everitt gives also a diagram of lines of 

 constant intensity). 



Ultramicroscopy . 



The considerations of this paper give no indication of the visi- 

 bility of isolated particles, but only of the possibility of detecting 

 their shape. If each gives sufficient light (either by self-luminosity, 

 as in the case of stars, or by illumination by a powerful beam 

 athwart the line of vision, as in ultramicroscopy), it will be seen. 

 The amount of light it scatters is proportional to the sixth power 

 of its radius when it is small compared with the wave-length. Its 

 image is almost independent of its shape under the same condition. 

 Under strong illumination larger particles ( <X) give complicated 

 diffraction figures; but not much can be learned from attempts to 

 interpret them. The visible disc is certainly much larger than the 

 geometrical image of the particle. Similarly, a luminous line gives 

 an image much wider than its geometric image. This case and that 

 of an isolated dark line of finite width on a bright background have 

 been worked out by Lord Rayleigh. In the latter case, when the 

 background consists of light all in one phase, he concludes that 

 the bar might w^ell remain visible when the width of the bar is 

 only one thirty-second part of the minimum distance between two 

 lines for resolution. The slightly darkened image of the bar has 

 then a width equal to about sixteen times that of its geometrical 

 image and its apparent width is therefore quite illusory. In the 

 case of a self-luminous background {i.e., with phases completely 

 independent), a bar of the same width has only half the visibility 

 of the previous case, but it should bs easily recognisable when its 



