dyed films of gelatin which transmitted a band in the extreme red King's College, Cambridge.

Transcription

1 PHYSIOLOGICAL ASPECT OF PHOTOGRAPHIC SAFE LIGHT SCREENS. BY H. HARTRIDGE, Fellow of King's College, Cambridge. (From the Physiological Laboratory, Cambridge.) DURING some experiments on a safe light screen for photographic purposes, several phenomena have been encountered of physiological interest. It had been found by means of a simple spectrophotometer that the majority of non-colour sensitive silver emulsions do not react to the part of the spectrum of longer wave-length than A 530, while several were found to be inactive to A 520. A few emulsions were found to be sensitive again in the extreme red of the spectrum but this was negligible in amount compared to that found in the blue, violet and ultra violet. Experiments were therefore made to ascertain what combination of waves above A 530 produced the best visual effect. Now in order to obtain a white light it would be necessary to combine with the extreme red of the spectrum, a band in the green complementary to it in colour. This was shown by Helmholtz(l) to be at about A 490 and it was therefore within the region to which a photographic emulsion is sensitive. A filter was made however by combining separate dyed films of gelatin which transmitted a band in the extreme red A 730 to A 685 and also one as low in the green as was safe, namely, from A 550 to A 530, in order to investigate the light passed by it. When this colour filter was placed in front of an osram lamp, the light transmitted was seen to be of a pale rose colour. When other lights in the dark room were extinguished the eye slowly became adapted to the colour of the light which then appeared more or less white(2). Particularly was this the case if the illumination was diminished, for the Purkinje effect(3) caused the green component to appear of relatively greater intensity than the red and therefore the two components more nearly balanced. When coloured objects were examined it was found that a considerable number could be recognized. Red, pink and brown, green, pale green and olive and various shades of grey appeared to be

2 96 H. HARTRIDGE. changed but little; certain greens however and most shades of blue and violet appeared red. The greens most noticeably affected were those of plants(4), the pale yellow-green changing to an autumnal brown in this light. The explanation was found to be that these greens, blues and violets all reflect brilliantly a band in the red, that almost exactly corresponded to that transmitted by the colour screen. There were two other noticeable features about the light: the first was that its colour depended to a great extent on the illuminant in front of which the filter was placed; thus by candle light it was red, while by daylight it was apple green (5). This phenomenon was due to the varying amounts of red and green present in the lights from different illuminants. The second feature was that if white objects on dark grounds were examined colour fringes were to be seen round their edges, for the eye being uncorrected for colour the red and green components came to a different focus on the retina and therefore if one was in focus it formed a sharp image which was overlapped by the confusion disk of the other. This also led to what may be termed chromatic instability, for some moments the red image would be focussed, the green forming the fringe, and then suddenly the green would be the focussed and the red would form the unfocussed image. Since the separation of the two component bands in the spectrum is the cause of the colour fringes, the chromatic instability, and the change in colour of the light with change in the artificial light source, while the use of a band in the extreme red in the spectrum gives the false colour rendering of certain greens, blues and violets, it was clear that the use of a different band in the red, one closer to the green band in the spectrum and one which greens, blues and violets do not reflect, would probably improve the visual effect of the light. Two objections to the change were to be anticipated, however, for the light would probably be less white owing to the fact that the colours would be still nearer together in the spectrum than they were before; there would also be difficulty in removing sharply the intermediate spectral region between the two bands. The experiment was tried however and proved successful. If the light from an osram lamp having passed through this filter was allowed to illuminate a white surface, then the appearance of the surface except for difference in intensity was almost indistinguishable from its appearance under the unscreened light. Careful comparison in a photometer showed that the surface was a pale cream when illuminated by the light that had passed through the filter. If the surface of the filter were itself looked at its colour was seen

3 SAFE LIGHT SCREENS. to be a yellowish green, which varied slightly in colour from part to part; this was due to the uneven dyeing of the gelatin films of which the filter was built up. By daylight the filter appeared a strong lemon yellow, but here again the light quickly lost its colour if direct daylight were excluded. The interposition of neutral tint glasses between the eye and the surface of the colour screen caused the latter to appear whiter. Any other method of reducing the intensity such as putting resistance into the osram lamp circuit brought about the same result. This light was found very pleasant to the eyes, no colour fringes could be observed round black and white objects, and there was no sign of chromatic instability. The colour of the transmitted light with different illuminants remained practically constant. Colour rendering of objects was good. No obvious change occurred in colours when the screen was interposed between the objects and the illuminant. Chlorophyl green even was found to be correctly represented, blues and violets were easily recognized. Comparison in a simple photometer showed that while yellows were too pale, blue greens and blues lacked brilliance while violets were too red. Spectroscopically the light consisted of a single broad spectral band extending from A 680 to A 530. At about the position of the yellow A 575 to A 595 there was a weak and indefinite absorption band which spread into the red and green. The region from A 680 to A 595 therefore appeared a full red of almost uniform colour; similarly a full green was seen to occupy the region A 575 to A 530. Between A 575 and A 595 the red and green areas fused into a brownish band. The orange and vellow green appeared to be absent from the spectrum. This was doubtless partly due to simultaneous contrast between the red and green areas, the orange appearing redder and the yellow green greener; partly also tc reduced intensity, for E dridge G r e en found that for normal vision there were fewer definite coloured areas seen in the spectrum with low intensity than there were for high (6). This was found to be so in this case, for on increasing the power of the light source or opening the slit of the spectroscope, the spectrum was seen to return to its normal appearance. It seemed remarkable that a light corresponding to a broad spectral area occupying roughly one-half the spectrum should under any circumstances appear white. Yet such is the case and the explanation appears to be as follows: Rayleigh has shown(7) that the yellow formed by mixing red and green of the right intensity closely matched a pure spectral yellow, the amount of red and green and white required for a match depending on the individual. PH. L. 7 97

4 98 H. HARTRIDGE. Between the extreme limit of the red and A 540 it is said that no white light is required for a perfect match. Thus v. Kries found(8) that red A and green A 552 when mixed in right amount could match exactly all intermediate waves in the spectrum. If therefore the whole of the spectrum above A 540 be transmitted by a colour filter one would expect the resulting mixture to be a pure yellow, if the relative intensities of the various red and green components be suitably adjusted. But since v. Kries has proved that the relationship between the colours is such that from the point of view of colour mixture they appear to lie on a straight line it is sufficient to have adjusted the whole red intensity against the whole green. Now if the intensity is high, such is indeed found experimentally to be the case, the light obtained is a lemon yellow which matches a pure yellow exactly. If the intensity is reduced however it is found that the miatch no longer holds, for the mixed yellow now has a white valency which is not present in the other. Now since the mixed light itself contains some yellow rays, limiting the band on one side to A 550 instead of A 540 reduces the white valenev for a particular intensity, similarly limiting the band to A 530 increases the white valency, while removing the central part of the band still further increases the whiteness of the transmitted light. Now as mentioned above the colour screen made by me transmits as far as A 530 and contains a dye that reduces as far as possible the intensity of the yellow rays. The best amount of reduction in practice was found to be such that it made the intensity of the yellow about one-third that of the red and green. If therefore we assume that red and green rays of certain wave-lengths combine to give a light, which has a white valency, which white valency increases in amount relative to that of colour as the intensity is diminished, we have an explanation of the appearances of the light and the colour screen under the circumstances described above. This question will be returned to shortly. There is a further point to account for, namely, how it is possible to recognize blue objects in the complete absence of blue rays. The presence of blue rays can be excluded for even a trace would at once show itself by rendering the light unsafe photographically. The explanation appears to be as follows: The colour of a pigment is complementary to the part of the spectrum occupied by its absorption band, a blue pigment, therefore, is one with an absorption band in the yellow. A blue object should therefore be recognizable in any light from which yellow rays can be removed by absorption,

5 SAFE LIGHT SCREENS. but only if the remaining rays combine to give something other than yellow. If therefore v. Kries is right and the spectrum from the extreme red to A 540 indeed falls along a straight line, then it would be impossible to explain the recognition of a blue object in such a light, but if as I have suggested above the light contains a white valency formed by the union of red and green rays, then the removal of yellow rays by the absorption band of the pigment renders the light reflected by it less yellow, that is the pigment appears blue. The hypothesis of v. Kries would appear therefore to require revision for, if the straight line relationship holds good at high and medium intensities, it apparently does not do so at low. Now the idea has suggested itself that the white valency at low intensities is due to the light nearly reaching the threshold of the retinal apparatus for day vision, and therefore stimulating with relatively greater force the apparatus used at night. There are two reasons why this cannot be the explanation; firstly the intensity of the light is too high to- be near the threshold of day vision, since it is approximately equal to that of an ordinary orange or red safe light of the same photographic safety. Now it is well known that orange and red rays stimulate the apparatus of night vision to only a slight extent (9), the intensity must be therefore well above the threshold of day vision for otherwise no light would be visible. And what is true of orange and red is equally true of any other light of equal intensity. The second reason for rejecting the above explanation is that the colours of objects are recognized, which clearly shows that day vision is concerned. An alternative theory therefore suggests itself, namely, that the blue visual sensation is stimulated by rays at any rate longer than A 550 since such rays when suitably combined contain a white valency. Supposing for a moment that the limit to the blue sensation is at A 560, then the waves between A 560 and A 530 stimulate this sensation and therefore give the light its unsaturated appearance. It is more difficult to find an explailation of the increased whiteness of the light with decreased intensity, in fact so unlikely did it appear at first sight that pains were taken to prove that such was indeed the case. It was foun(d however that the observation of the phenomenon is by no means limited to this case. Thus Ton n (lo) found that although green A was complementary to red A 670-8, yet when the intensity was diminished the complementary green required was nearer the yellow, namely, A Similarly Aubert(11) found that a mixture

6 100 H. HARTRIDGE. of red and green that matched a pure yellow at high intensities, became relatively brighter and less saturated as the illumination was decreased. These changes are due according to Tschermak (12) to alteration in adaptation, the blue sensation becoming relatively more sensitive than the red and green with decreased illumination. The following experimental data would appear to be in favour of this view. If the eye be exposed to a strong yellow light for some minutes, and then the light from the colour screen examined it will be found that the light appears pure white even if the intensity is such that normally it would appear yellow. Similarly by exposing to blue rays, the light appears a pure lemon yellow even when of such low intensity that normally it would appear white. SUMMARY. It has been found by experiment that a photographic safe light screen can be constructed which gives better visual results than the orange and red screens usually employed. The colour of the light is found to vary with the inten-sity of the light, being yellow at high, cream at medium, and white at low intensities. Spectroscopically the light shows a broad band corresponding to the red, orange, yellow and yellowish green parts of the spectrum. The long red rays and the blue green, blue and ultra violet rays are not present. The colours of objects are only slightly modified by this light, the most noticeable change being that yellows are too pale, while blues are too dark. The white appearance of the light when of low intensity and the recognition of blue colouts in the absence of blue light can be explained on the tri-chromatic theory if one supposes that the blue sensation is stimulated by rays longer than A 550, and that this sensation is relatively more sensitive to light of low intensity than are the red and green sensations. REFERENCES. (1) Helmholtz. 3rd edit. II. p (2) Briicke. Ann. d. Phys. u. Chem. LXXXIV. p (3) HRring. Arch. f. d. ges. Physiol. LX. p (4) Rood. Modern Chromatics, p. 84. (5) Rayleigh. Collected papers, i. p. 79. (6) Edridge Green. Trans. Ophthal. Soc. XXVII. p. 2. (7) Rayleigh. Collected papers, i. p (8) v. Kries. Ztschr. f. Psychol. u. Physiol. d. Sinnesorg. XIII. p (9) Abney. Phil. Trans. CLXXXIII. p (10) Tonn. Ztschr. f. Psychol. u. Physiol. d. Sinnesorg. via. p (11) Aubert. Weidemann's Ann. XVI. p (12) Tschermalk. Arch. f. d. ges. Physiol. LXX. p