HENRY M. SAYERS 167 



Similarly, if a cover-glass type of reflector is used the maximum 

 illumination possible theoretically would' be given if the reflector 

 reflected 50 per cent, of the light and transmitted 50 per cent. The 

 result would be 25 per cent, of the light utilised. No cover-glass 

 type reflector does nearly so well as this. Measurements of a clear 

 cover-glass have given me a reflective power of 13.8 per cent, com- 

 pared with a silvered total reflection prism, and a transmission of 

 66 per cent., with the glass inclined at 45° to the beam. The 

 resultant for the image illumination is therefore 13.8 x 66 per cent, 

 or 9.2 per cent, for the image. A semi-platinised cover-glass (not 

 made for the purpose) gave 38 per cent, reflection, but only 21.5 per 

 cent, transmission or 8.17 per cent, for the image. The light, too, 

 was brownish yellow. 



The prism reflector, while much superior to the cover-glass in 

 respect to illumination, cuts down the effective aperture of the objec- 

 tive, both as a condenser and as an objective, and does this unsym- 

 metrically to the detriment of its resolving power. This is probably 

 the reason why prism reflectors are never made to cover nearly so 

 much as half the objective aperture, and they are consequently not in 

 practice much better than cover-glasses in respect to brightness of 

 image. 



The diminution of light intensity in the image compared with that 

 reflected by the object is proportional to the square of the lineal magni- 

 fication, increased by absorption and reflection in the optical system. 

 With 100 diameters the light intensity at the image is under one ten- 

 thousandth, and with 1,000 diameters under one-millionth part of the 

 light reflected by the object. Allowing for the loss due to the vertical 

 reflector it is for the two magnifications mentioned less than one forty- 

 thousandth and one four-millionth part respectively of the light inten- 

 sity of the source for any part of the object which has perfect reflecting 

 power, assuming that the image of the source is as bright as the source 

 itself. As these small fractions are on assumptions of 100 per cent, 

 efficiency for every stage of reflection and refraction between the 

 source and the image, which is unattainable everywhere, it may be 

 roughly estimated that the actual fractions of the source brightness in 

 the brightest points of the image will be more nearly one ten-thousandth 

 and one ten-millionth respectively. The most obvious improvement 

 would be the devising of a vertical reflector giving an efficiency of com- 

 bined reflection and transmission approximating to 25 per cent, without 

 cutting down the objective aperture. 



The exposure required is that for the " shadows " of the object, 

 i.e., for the darkest parts which show perceptible' detail. It follows 

 that to get reasonable exposure times only light sources of great 

 intrinsic brilliancy are of practical use for photographic work. The 

 total candle-power of the radiant is no criterion by itself, it is candle- 

 power per unit area of radiating surface which counts. 



Of the available sources the positive crater of the carbon arc is 

 the most brilliant. After that come in descending order the Nernst 

 lamp, the tungsten arc (or Pointolite), the half-watt metal filament 

 lamp, and the oxy-hydrogen lime-light. The last named is the only 

 light source depending on combustion at all suitable for the purpose, 



