68 



SCIENCE- G OS SIP. 



repeated measurements of the same glass. The 

 sectors are revolving at about 4,000 revolutions per 

 minute, and it is only when the motor driving them 

 has been stopped that we can read their aperture , 

 so that it is impossible for the observer to be 

 influenced in his estimation by his knowledge of 

 previous readings ; yet twenty or thirty successive 

 readings of the same glass will not vary more than 

 1 per cent. By this instrument we measure once 

 for all the brightness of the light transmitted by 

 each of our test glasses as against white light, and 

 from these data it is easy to calculate what pro- 

 portion of light the opacities of our negatives of 

 the test object should transmit. 



Sir William Abney has also given us the means 

 whereby we may avoid the measurement of each 

 negative. To do this we mount in front of the row 

 of glasses a revolving sector with apertures so 

 arranged that the light coming through each glass 

 of the series may be reduced to equal luminosity — 

 the luminosity of the darkest glass in the series. 

 "When once we have this instrument set up, the 

 testing of a light filter is a very easy matter. A 

 photograph taken through the filter under test 

 should give us a row of patches of equal density, 

 and as the eye is a very excellent judge of the 

 quality of the two adjacent opacities, we can tell 

 at once by simple observation the accuracy of our 

 light filter. 



With regard to the light filter itself, in addition 

 to its power of correcting accurately the imper- 

 fections of the photographic plate, there are two 

 conditions which it must fulfil. First, it must have 

 a fair amount of permanency and durability, and, 

 secondly, it must not affect the definition of the 

 lens. Filters might be made by filling a glass cell 

 with a coloured liquid, but such an arrangement is 

 not very satisfactory in practice, especially if used 

 cut of doors, on account of the risk of accidental 

 damage or leakage, and the bad effect on definition 

 caused by the circulation of the liquid with every 

 change of temperature. Therefore, for commercial 

 23urposes, we are restricted to a coloured film, sealed 

 in optical contact between two pieces of optically 

 ground and polished plate glass. Fortunately, the 

 aniline dyes give us an almost endless variety of 

 absorptions, and a large number of these which 

 have been carefully selected and tested prove to be 

 quite permanent, when protected by being sealed 

 between two glass plates. By using combinations 

 of these dyes in varying depths of tint, it is possible 

 to adjust a filter to almost any curve of the spec- 

 trum. In the early days of orthochromatic photo- 

 graphy a piece of ordinary yellow glass, ground and 

 polished, was used as a substitute for a correct 

 light filter ; such glasses, however, cut out the 

 action of the blue entirely, a defect almost as bad 

 as the evil it was supposed to avoid. 



Some years ago I showed at the exhibition of 

 the Eoyal Photographic Society a frame of dia- 

 phragm light filters, in which the colour film was 



pierced in the centre, so that a small portion of 

 white light would be mixed with a far larger pro- 

 portion of the filtered light. This, however, was 

 a clumsy device and in the 'later filters has been 

 done away with, by using two or more dyed films, 

 so that one dye would reduce the action of the 

 green, and another dye the action of the blue. By 

 respectively varying the depth of tint of these colour 

 films, we can correct for definite regions of the 

 spectrum, and it is only by this device that the 

 production of filters of great accuracy has been 

 made commercially possible. For instance, the 

 Cadett light filter gives practically perfect ren- 

 dering of all colours when used with the Cadett 

 rapid spectrum plate, and yet these filters are but 

 little more expensive than the old pot -yellow glass. 

 An examination of a number of the Cadett light 

 filters and comparison with the photographic test 

 plates reveals several interesting points. For one 

 thing it proves how utterly unreliable is the eye 

 when used as a judge of a light filter. In reality 

 the light filter is cutting out what the eye does not 

 see, the photographic plate being sensitive to 

 luminosities of blue and violet so low as to be 

 quite invisible to the human eye. Perhaps half-a- 

 dozen niters will appear exactly the same to the 

 eye, although differing very widely indeed in their 

 filtering power, as shown by the deposits in the 

 respective patches of the test negatives. On the 

 other hand two niters, the differences in the colour 

 of which are distinctly visible to the eye, may 

 exactly correspond in their filtering power as shown 

 by the test negative. 



When we compare the very wide difference 

 between the curve given by the spectrum on the 

 Cadett lightning spectrum plate without a filter 

 (Xo. 7, fig. 4) with the ideal curve (fig. 2), we see 

 that the effect of a filter which will sufficiently 

 reduce the action of the blue, so as to allow the 

 reds and greens to impress the plate correctly, must 

 considerably increase the time of exposure neces- 

 sary for securing a negative. With the Cadett 

 " Absolutus " filter, exposures are increased about 

 forty times, but owing to the extreme speed of the 

 plate, far in advance of any other sensitive surface 

 known, the exposures in the studio are compara- 

 tively short, and negatives of brightly-lighted land- 

 scapes fully exposed may be obtained in a fraction 

 of a second. 



Since the " Absolutus " filter was placed upon 

 the market Messrs. Cadett have introduced another 

 filter, which they term " Gilvus."' This filter is 

 made in precisely the same accurate manner as the 

 " Absolutus " filter, with regard to its rendering of 

 blue and green, but is not corrected for red. By 

 ignoring the red the filter can be made much more 

 transparent, so that the increased time of exposure 

 is only about four times that required without the 

 filter. This filter proves very valuable in landscape 

 work, and saves a great deal of retouching in 

 portraiture : and where the object photographed 



